{"meta": {"count": 1000, "db_response_time_ms": 34, "page": 1, "per_page": 100, "groups_count": null, "x_query": {"oql": "get works where retracted is (false)\n  and year is (2025)\n  and title-abstract has ((stemmed \"language model\" OR stemmed \"large language model\" OR stemmed \"vision language model\") AND (stemmed \"linear attention\" OR stemmed \"long context\" OR stemmed \"memory augmented\" OR stemmed \"mixture of experts\" OR stemmed \"state space model\" OR stemmed \"transformer architecture\"))\n  and type is (article or preprint or review);\nthen sample (1000) of those works with seed (42)", "oqo": {"get_rows": "works", "filter_rows": [{"column_id": "is_retracted", "value": false}, {"column_id": "publication_year", "value": 2025}, {"join": "or", "filters": [{"column_id": "title_and_abstract.search", "value": "\"language model\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"large language model\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"vision language model\"", "operator": "has"}]}, {"join": "or", "filters": [{"column_id": "title_and_abstract.search", "value": "\"linear attention\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"long context\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"memory augmented\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"mixture of experts\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"state space model\"", "operator": "has"}, {"column_id": "title_and_abstract.search", "value": "\"transformer architecture\"", "operator": "has"}]}, {"join": "or", "filters": [{"column_id": "type", "value": "article"}, {"column_id": "type", "value": "preprint"}, {"column_id": "type", "value": "review"}]}], "sample": 1000, "seed": "42"}, "url": "/works?filter=is_retracted:false,publication_year:2025,title_and_abstract.search:%22language model%22|%22large language model%22|%22vision language model%22,title_and_abstract.search:%22linear attention%22|%22long context%22|%22memory augmented%22|%22mixture of experts%22|%22state space model%22|%22transformer architecture%22,type:article|preprint|review&sample=1000&seed=42&per_page=100&page=1"}, "cost_usd": 0.001}, "results": [{"id": "https://openalex.org/W4405990119", "doi": "https://doi.org/10.2139/ssrn.5080078", "title": "Long Context Window-Based Legal Interpretation of Building Codes and Regulations", "publication_year": 2025, "publication_date": "2025-01-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Jaekun Lee", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I193775966", "display_name": "Yonsei University", "ror": "https://ror.org/01wjejq96", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I193775966"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Jaekun Lee", "raw_affiliation_strings": ["Yonsei University", "Yonsei University, Seoul, Korea, Republic of (South Korea)"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Yonsei University", "institution_ids": ["https://openalex.org/I193775966"]}, {"raw_affiliation_string": "Yonsei University, Seoul, Korea, Republic of (South Korea)", "institution_ids": ["https://openalex.org/I193775966"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5075755166", "display_name": "Ghang Lee", "orcid": "https://orcid.org/0000-0002-3522-2733", "observed_orcids": ["https://orcid.org/0000-0002-3522-2733"]}, "institutions": [{"id": "https://openalex.org/I193775966", "display_name": "Yonsei University", "ror": "https://ror.org/01wjejq96", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I193775966"]}], "countries": ["KR"], "is_corresponding": true, "raw_author_name": "Ghang Lee", "raw_affiliation_strings": ["Yonsei University", "Yonsei University, Seoul, Korea, Republic of (South Korea)"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Yonsei University", "institution_ids": ["https://openalex.org/I193775966"]}, {"raw_affiliation_string": "Yonsei University, Seoul, Korea, Republic of (South Korea)", "institution_ids": ["https://openalex.org/I193775966"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T13643", "display_name": "Artificial Intelligence in Law", "score": 0.9326000213623047, "subfield": {"id": "https://openalex.org/subfields/3320", "display_name": "Political Science and International Relations"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T13643", "display_name": "Artificial Intelligence in Law", "score": 0.9326000213623047, "subfield": {"id": "https://openalex.org/subfields/3320", "display_name": "Political Science and International Relations"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.05990000069141388, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10260", "display_name": "Software Engineering Research", "score": 0.0020000000949949026, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2139/ssrn.5080078", "is_oa": true, "landing_page_url": "https://doi.org/10.2139/ssrn.5080078", "pdf_url": null, "source": {"id": "https://openalex.org/S4210172589", "display_name": "SSRN Electronic Journal", "issn_l": "1556-5068", "issn": ["1556-5068"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I1318003438", "host_organization_name": "RELX Group (Netherlands)", "host_organization_lineage": ["https://openalex.org/I1318003438"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416059721", "doi": "https://doi.org/10.48550/arxiv.2506.13734", "title": "Instruction Following by Principled Boosting Attention of Large Language Models", "publication_year": 2025, "publication_date": "2025-06-16", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5033315494", "display_name": "Vit\u00f3ria Guardieiro", "orcid": "https://orcid.org/0000-0003-1956-5418", "observed_orcids": ["https://orcid.org/0000-0003-1956-5418"]}, "institutions": [{"id": "https://openalex.org/I79576946", "display_name": "University of Pennsylvania", "ror": "https://ror.org/00b30xv10", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I79576946"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Guardieiro, Vitoria", "raw_affiliation_strings": ["University of Pennsylvania"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Pennsylvania", "institution_ids": ["https://openalex.org/I79576946"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Khare, Avishree", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I79576946", "display_name": "University of Pennsylvania", "ror": "https://ror.org/00b30xv10", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I79576946"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Khare, Avishree", "raw_affiliation_strings": ["University of Pennsylvania"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Pennsylvania", "institution_ids": ["https://openalex.org/I79576946"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Stein, Adam", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I79576946", "display_name": "University of Pennsylvania", "ror": "https://ror.org/00b30xv10", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I79576946"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Stein, Adam", "raw_affiliation_strings": ["University of Pennsylvania"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Pennsylvania", "institution_ids": ["https://openalex.org/I79576946"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5066376294", "display_name": "Eric Wong", "orcid": "https://orcid.org/0000-0002-8568-6659", "observed_orcids": ["https://orcid.org/0000-0002-8568-6659"]}, "institutions": [{"id": "https://openalex.org/I79576946", "display_name": "University of Pennsylvania", "ror": "https://ror.org/00b30xv10", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I79576946"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Wong, Eric", "raw_affiliation_strings": ["University of Pennsylvania"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Pennsylvania", "institution_ids": ["https://openalex.org/I79576946"]}]}], "abstract_inverted_index": {"Large": [0], "language": [1], "models'": [2], "behavior": [3], "is": [4, 62], "often": [5], "shaped": [6], "by": [7, 83, 136], "instructions": [8], "such": [9, 60], "as": [10, 87], "system": [11], "prompts,": [12], "refusal": [13], "boundaries,": [14], "privacy": [15], "constraints,": [16], "and": [17, 46, 93, 161, 170, 192], "tool-use": [18], "rules": [19, 92, 100], "that": [20, 53, 104, 128, 148], "must": [21], "hold": [22], "at": [23], "inference": [24], "time.": [25], "Yet": [26], "in": [27], "practice": [28], "these": [29], "constraints": [30], "can": [31, 124], "be": [32, 130], "violated": [33], "under": [34], "long": [35], "contexts": [36], "or": [37, 180], "when": [38], "user-provided": [39], "context": [40, 117, 127], "conflicts": [41], "with": [42, 96], "them,": [43], "creating": [44], "reliability": [45], "safety": [47], "risks.": [48], "This": [49], "motivates": [50], "inference-time": [51], "interventions": [52], "strengthen": [54], "instruction": [55, 69, 85, 91, 108, 194], "influence": [56], "without": [57], "retraining.": [58], "One": [59], "intervention": [61, 147], "attention": [63, 67, 80, 97, 106, 156, 172, 198], "steering,": [64, 169], "which": [65, 99], "biases": [66], "toward": [68], "tokens.": [70], "In": [71], "this": [72, 111, 137], "work,": [73], "we": [74, 139], "present": [75], "a": [76, 145, 150, 201], "unifying": [77], "theory": [78], "for": [79, 116], "steering": [81, 173], "methods": [82, 174, 191], "formalizing": [84], "following": [86], "rule-based": [88], "competition": [89], "between": [90], "context-derived": [94], "rules,": [95], "mediating": [98], "dominate.": [101], "We": [102, 163], "prove": [103], "boosting": [105, 123], "to": [107, 118, 154], "tokens": [109], "tilts": [110], "competition,": [112], "making": [113], "it": [114], "harder": [115], "override": [119], "instruction-following.": [120], "However,": [121], "excessive": [122], "suppress": [125], "task-relevant": [126], "should": [129], "incorporated": [131], "alongside": [132], "the": [133, 186, 193], "instruction.": [134], "Guided": [135], "theory,": [138], "propose": [140], "Instruction": [141], "Attention": [142], "Boosting": [143], "(InstABoost),": [144], "simple": [146], "applies": [149], "constant": [151], "additive": [152], "bias": [153], "instruction-key": [155], "logits": [157], "across": [158, 175], "all": [159, 182], "layers": [160], "heads.": [162], "evaluate": [164], "InstABoost": [165, 178], "against": [166], "prompting,": [167], "latent": [168, 190], "prior": [171, 197], "15": [176], "tasks.": [177], "matches": [179], "outperforms": [181], "baselines": [183], "while": [184], "avoiding": [185], "fluency": [187], "collapse": [188], "of": [189, 196], "over-focus": [195], "methods,": [199], "achieving": [200], "stronger": [202], "steering-quality": [203], "tradeoff.": [204]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7807999849319458, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7807999849319458, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.14900000393390656, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.014999999664723873, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.13734", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.13734", "pdf_url": "https://arxiv.org/pdf/2506.13734", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4412127082", "doi": "https://doi.org/10.2139/ssrn.5333028", "title": "Can Memory-Augmented Large Language Model Agents Help Journalism in News Interpretation and Cognitive Framing Across Diverse Audiences?", "publication_year": 2025, "publication_date": "2025-01-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5117935403", "display_name": "Leyi Ouyang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I101479585", "display_name": "South China Agricultural University", "ror": "https://ror.org/05v9jqt67", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I101479585"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Leyi Ouyang", "raw_affiliation_strings": ["South China Agriculture University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "South China Agriculture University", "institution_ids": ["https://openalex.org/I101479585"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.28610000014305115, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.28610000014305115, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10133", "display_name": "Media Studies and Communication", "score": 0.17749999463558197, "subfield": {"id": "https://openalex.org/subfields/3315", "display_name": "Communication"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T13629", "display_name": "Text Readability and Simplification", "score": 0.07779999822378159, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2139/ssrn.5333028", "is_oa": true, "landing_page_url": "https://doi.org/10.2139/ssrn.5333028", "pdf_url": null, "source": {"id": "https://openalex.org/S4210172589", "display_name": "SSRN Electronic Journal", "issn_l": "1556-5068", "issn": ["1556-5068"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I1318003438", "host_organization_name": "RELX Group (Netherlands)", "host_organization_lineage": ["https://openalex.org/I1318003438"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7118317836", "doi": "https://doi.org/10.15680/ijirset.2025.1412113", "title": "The Symbiotic Evolution: Modern AI Algorithms and the Paradigm Shift to DataCentric Technologies", "publication_year": 2025, "publication_date": "2025-01-01", "type": "article", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5142582759", "display_name": "Sahaya Brainy", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405271358", "display_name": "Jerusalem College of Engineering", "ror": "https://ror.org/02aae6a21", "country_code": "IN", "type": "education", "lineage": ["https://openalex.org/I4405271358"]}], "countries": ["IN"], "is_corresponding": true, "raw_author_name": "Sahaya Brainy", "raw_affiliation_strings": ["Assistant Professor, Dept. of Artificial Intelligence and Data Science , Jerusalem College of Engineering , Pallikaranai , Chennai , India"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Assistant Professor, Dept. of Artificial Intelligence and Data Science , Jerusalem College of Engineering , Pallikaranai , Chennai , India", "institution_ids": ["https://openalex.org/I4405271358"]}]}], "abstract_inverted_index": {"ABSTRACT:": [0], "The": [1, 283, 300], "landscape": [2], "of": [3, 51, 74, 103, 156, 258, 289, 325, 354], "Artificial": [4], "Intelligence": [5], "(AI)": [6], "is": [7, 31, 133, 292], "undergoing": [8], "a": [9, 19, 48, 164, 170, 185, 201, 205, 209, 230, 243, 255, 266], "fundamental": [10], "reorientation.": [11], "While": [12], "the": [13, 35, 52, 60, 72, 79, 101, 134, 154, 197, 219, 287, 351], "past": [14], "decade": [15], "was": [16], "defined": [17], "by": [18, 275, 281, 316, 320], "\"model-centric\"": [20], "approach\u2014focusing": [21], "on": [22, 228], "architectural": [23], "innovations": [24], "in": [25, 151, 234, 251, 297], "neural": [26, 93], "networks\u2014a": [27], "compelling": [28], "\"data-centric\"": [29], "paradigm": [30], "now": [32], "emerging": [33], "as": [34, 350], "critical": [36], "frontier": [37], "for": [38, 242, 269, 340], "robust,": [39], "scalable,": [40], "and": [41, 59, 65, 82, 91, 124, 159, 177, 183, 208, 277, 307, 329, 346], "trustworthy": [42], "AI": [43, 57, 76, 126, 261, 290], "systems.This": [44], "research": [45, 335], "paper": [46], "presents": [47], "comprehensive": [49], "analysis": [50, 284], "symbiotic": [53], "relationship": [54], "between": [55, 200], "modern": [56], "algorithms": [58, 143, 158], "data-centric": [61, 104, 125, 225, 298], "technologies": [62], "that": [63, 224, 264, 286], "enable": [64], "amplify": [66], "their": [67], "effectiveness.": [68], "We": [69, 181, 332], "first": [70], "delineate": [71], "evolution": [73], "core": [75], "algorithms,": [77], "from": [78, 218], "Transformer": [80], "architecture": [81], "large": [83], "language": [84], "models": [85, 90, 140, 339], "(LLMs)": [86], "to": [87, 195, 238, 295, 302], "multimodal": [88], "foundation": [89, 338], "efficient": [92], "architectures": [94], "like": [95], "mixture-of-experts": [96], "(MoE).": [97], "Concurrently,": [98], "we": [99], "map": [100], "ecosystem": [102], "technologies,": [105], "encompassing": [106], "advanced": [107], "data": [108, 112, 115, 122, 146, 160, 215, 322, 341, 344, 348], "engineering": [109], "(vector": [110], "databases,": [111], "lakes),": [113], "automated": [114], "preparation": [116], "(data": [117], "programming,": [118], "weak": [119], "supervision),": [120], "synthetic": [121], "generation,": [123], "operations": [127], "(DataOps,": [128], "MLOps).": [129], "A": [130], "central": [131], "contribution": [132], "\"Algorithm-Data": [135], "Virtuous": [136], "Cycle\"": [137], "framework,": [138], "which": [139, 150], "how": [141], "sophisticated": [142], "unlock": [144], "richer": [145], "representations": [147], "(e.g.,": [148], "embeddings),": [149], "turn": [152], "fuel": [153], "development": [155], "next-generation": [157], "management": [161, 271], "tools.": [162], "Employing": [163], "multi-method": [165], "approach,": [166], "this": [167], "study": [168, 257], "combines": [169], "systematic": [171], "literature": [172], "review": [173], "with": [174, 247], "quantitative": [175], "experiments": [176], "qualitative": [178], "case": [179, 256], "analysis.": [180], "designed": [182], "executed": [184], "controlled": [186], "experiment": [187], "across": [188], "three": [189], "domains": [190], "(computer": [191], "vision,": [192], "NLP,": [193], "time-series)": [194], "quantify": [196], "performance": [198], "delta": [199], "model-centric": [202, 240], "optimization": [203, 211], "(tuning": [204], "state-of-the-art": [206], "model)": [207], "datacentric": [210], "(systematically": [212], "improving": [213], "training": [214], "quality)": [216], "starting": [217], "same": [220], "baseline.": [221], "Results": [222], "demonstrated": [223], "interventions": [226], "yielded,": [227], "average,": [229], "15.8%": [231], "greater": [232], "improvement": [233], "model": [235], "accuracy": [236, 280], "compared": [237], "additional": [239], "tuning": [241], "fixed": [244], "compute": [245], "budget,": [246], "gains": [248], "exceeding": [249], "25%": [250], "low-data": [252], "regimes.": [253], "Furthermore,": [254], "an": [259], "industrial": [260], "pipeline": [262], "revealed": [263], "implementing": [265], "vector": [267], "database": [268], "embedding": [270], "reduced": [272], "inference": [273], "latency": [274], "40%": [276], "improved": [278], "retrieval": [279], "18%.": [282], "concludes": [285], "future": [288], "progress": [291], "inextricably": [293], "linked": [294], "advancements": [296], "technologies.": [299], "path": [301], "artificial": [303], "general": [304], "intelligence": [305], "(AGI)": [306], "reliable": [308], "real-world": [309], "deployment": [310], "will": [311], "be": [312], "paved": [313], "not": [314], "merely": [315], "larger": [317], "models,": [318], "but": [319], "smarter": [321], "systems": [323], "capable": [324], "curation,": [326, 345], "synthesis,": [327], "validation,": [328], "continuous": [330], "evolution.": [331], "identify": [333], "key": [334], "vectors": [336], "including": [337], "tasks,": [342], "causal": [343], "federated": [347], "ecosystems": [349], "next": [352], "pillars": [353], "advancement.": [355]}, "primary_topic": {"id": "https://openalex.org/T12535", "display_name": "Machine Learning and Data Classification", "score": 0.4011000096797943, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T12535", "display_name": "Machine Learning and Data Classification", "score": 0.4011000096797943, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.3041999936103821, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.03680000081658363, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.15680/ijirset.2025.1412113", "is_oa": true, "landing_page_url": "https://doi.org/10.15680/ijirset.2025.1412113", "pdf_url": "https://doi.org/10.15680/ijirset.2025.1412113", "source": {"id": "https://openalex.org/S2764936073", "display_name": "International Journal of Innovative Research in Science Engineering and Technology", "issn_l": "2319-8753", "issn": ["2319-8753", "2347-6710"], "is_oa": true, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core"], "host_organization": "https://openalex.org/P4323238129", "host_organization_name": "Ess And Ess Research Publications", "host_organization_lineage": ["https://openalex.org/P4323238129"], "host_organization_lineage_names": ["Ess And Ess Research Publications"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "International Journal of Innovative Research in Science Engineering and Technology", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4408184554", "doi": "https://doi.org/10.63337/term.2025.53588", "title": "The Long Context Conundrum: Challenges and Innovations in Scaling LLM Memory", "publication_year": 2025, "publication_date": "2025-03-03", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5101381944", "display_name": "Anjanava Biswas", "orcid": "https://orcid.org/0009-0008-3225-3316", "observed_orcids": ["https://orcid.org/0009-0008-3225-3316"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Anjanava Biswas", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0008-3225-3316", "affiliations": []}], "abstract_inverted_index": {"In": [0], "an": [1, 89, 113], "era": [2], "where": [3], "AI": [4, 29, 151], "assistants": [5], "are": [6, 31], "expected": [7], "to": [8, 33, 48, 54], "digest": [9], "entire": [10], "novels,": [11], "multi-document": [12], "dossiers,": [13], "or": [14, 96], "hours": [15, 97], "of": [16, 76, 94, 98, 145], "conversation,": [17], "the": [18, 146], "demand": [19], "for": [20], "longer-context": [21], "large": [22], "language": [23], "models": [24, 62], "(LLMs)": [25], "is": [26, 69, 141], "soaring.": [27], "Top": [28], "companies": [30], "racing": [32], "expand": [34], "context": [35, 107, 127], "windows:": [36], "OpenAI\u2019s": [37], "GPT-4": [38], "introduced": [39], "32,000-token": [40], "versions,": [41], "Anthropic\u2019s": [42], "Claude": [43], "leapt": [44], "from": [45], "a": [46, 121, 129], "9K": [47], "100K": [49], "token": [50, 125], "window": [51], "(and": [52], "later": [53], "200K),": [55], "and": [56, 135], "startups": [57], "like": [58], "AI21": [59], "have": [60], "unveiled": [61], "with": [63, 74, 105], "256K-token": [64], "capacity": [65], "[1].": [66, 84], "Google": [67], "DeepMind": [68], "reportedly": [70], "aiming": [71], "even": [72], "higher,": [73], "hints": [75], "million-token": [77], "contexts": [78], "in": [79, 101, 126, 132, 149], "its": [80], "upcoming": [81], "Gemini": [82], "model": [83], "These": [85], "numbers": [86], "promise": [87], "that": [88], "LLM": [90], "could": [91], "ingest": [92], "hundreds": [93], "pages": [95], "transcribed": [99], "audio": [100], "one": [102, 144], "go.": [103], "But": [104], "great": [106, 109], "comes": [108], "computation": [110], "\u2013": [111], "extending": [112], "LLM\u2019s": [114], "memory": [115, 136], "isn\u2019t": [116], "as": [117, 119, 143], "simple": [118], "turning": [120], "dial.": [122], "Each": [123], "additional": [124], "exacts": [128], "steep": [130], "cost": [131], "processing": [133], "time": [134], "usage.": [137], "The": [138], "long-context": [139], "conundrum": [140], "emerging": [142], "toughest": [147], "challenges": [148], "scaling": [150], "capabilities.": [152]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8999000191688538, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8999000191688538, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.07209999859333038, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.0035000001080334187, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.63337/term.2025.53588", "is_oa": false, "landing_page_url": "https://doi.org/10.63337/term.2025.53588", "pdf_url": null, "source": null, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "The Edge Review", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4406450004", "doi": "https://doi.org/10.48550/arxiv.2501.08262", "title": "Addressing the sustainable AI trilemma: a case study on LLM agents and RAG", "publication_year": 2025, "publication_date": "2025-01-14", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5016274096", "display_name": "Hui Wu", "orcid": "https://orcid.org/0000-0003-3930-1225", "observed_orcids": ["https://orcid.org/0000-0003-3930-1225"]}, "institutions": [{"id": "https://openalex.org/I23923803", "display_name": "University of Exeter", "ror": "https://ror.org/03yghzc09", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I23923803"]}], "countries": ["GB"], "is_corresponding": false, "raw_author_name": "Wu, Hui", "raw_affiliation_strings": ["Dept. of Engineering University of Exeter Exeter , UK EX4 4QF"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Dept. of Engineering University of Exeter Exeter , UK EX4 4QF", "institution_ids": ["https://openalex.org/I23923803"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100334504", "display_name": "Xiaoyang Wang", "orcid": "https://orcid.org/0009-0001-0362-317X", "observed_orcids": ["https://orcid.org/0009-0001-0362-317X", "https://orcid.org/0000-0002-0650-6868"]}, "institutions": [{"id": "https://openalex.org/I23923803", "display_name": "University of Exeter", "ror": "https://ror.org/03yghzc09", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I23923803"]}], "countries": ["GB"], "is_corresponding": false, "raw_author_name": "Wang, Xiaoyang", "raw_affiliation_strings": ["Dept. of Computer Science University of Exeter Exeter , UK EX4 4QF"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Dept. of Computer Science University of Exeter Exeter , UK EX4 4QF", "institution_ids": ["https://openalex.org/I23923803"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5115902717", "display_name": "Zhong Fan", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I23923803", "display_name": "University of Exeter", "ror": "https://ror.org/03yghzc09", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I23923803"]}], "countries": ["GB"], "is_corresponding": false, "raw_author_name": "Fan, Zhong", "raw_affiliation_strings": ["Dept. of Engineering University of Exeter Exeter , UK EX4 4QF"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Dept. of Engineering University of Exeter Exeter , UK EX4 4QF", "institution_ids": ["https://openalex.org/I23923803"]}]}], "abstract_inverted_index": {"Large": [0], "language": [1], "models": [2], "(LLMs)": [3], "have": [4], "demonstrated": [5], "significant": [6, 85], "capabilities,": [7], "but": [8], "their": [9], "widespread": [10], "deployment": [11], "and": [12, 42, 53, 67, 78, 92, 111], "more": [13, 117], "advanced": [14], "applications": [15], "raise": [16], "critical": [17], "sustainability": [18], "challenges,": [19], "particularly": [20], "in": [21, 63, 88, 108], "inference": [22], "energy": [23, 60, 76, 86], "consumption.": [24], "We": [25], "propose": [26], "the": [27, 30, 35, 59, 73, 104], "concept": [28], "of": [29, 50], "Sustainable": [31], "AI": [32, 38, 119], "Trilemma,": [33], "highlighting": [34], "tensions": [36], "between": [37, 75], "capability,": [39], "digital": [40], "equity,": [41], "environmental": [43], "sustainability.": [44], "Through": [45], "a": [46], "systematic": [47], "case": [48], "study": [49], "LLM": [51], "agents": [52], "retrieval-augmented": [54], "generation": [55], "(RAG),": [56], "we": [57], "analyze": [58], "costs": [61], "embedded": [62], "memory": [64], "module": [65], "designs": [66], "introduce": [68], "novel": [69], "metrics": [70], "to": [71], "quantify": [72], "trade-offs": [74], "consumption": [77], "system": [79], "performance.": [80], "Our": [81, 101], "experimental": [82], "results": [83], "reveal": [84], "inefficiencies": [87], "current": [89], "memory-augmented": [90], "frameworks": [91], "demonstrate": [93], "that": [94], "resource-constrained": [95], "environments": [96], "face": [97], "disproportionate": [98], "efficiency": [99], "penalties.": [100], "findings": [102], "challenge": [103], "prevailing": [105], "LLM-centric": [106], "paradigm": [107], "agent": [109], "design": [110], "provide": [112], "practical": [113], "insights": [114], "for": [115], "developing": [116], "sustainable": [118], "systems.": [120]}, "primary_topic": {"id": "https://openalex.org/T12238", "display_name": "Green IT and Sustainability", "score": 0.9643999934196472, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T12238", "display_name": "Green IT and Sustainability", "score": 0.9643999934196472, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.01549999974668026, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T14098", "display_name": "Energy Efficiency in Computing", "score": 0.010300000198185444, "subfield": {"id": "https://openalex.org/subfields/1708", "display_name": "Hardware and Architecture"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2501.08262", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2501.08262", "pdf_url": "https://arxiv.org/pdf/2501.08262", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4407123511", "doi": "https://doi.org/10.48550/arxiv.2502.00344", "title": "FinchGPT: a Transformer based language model for birdsong analysis", "publication_year": 2025, "publication_date": "2025-02-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Kobayashi, Kosei", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I201537933", "display_name": "Tohoku University", "ror": "https://ror.org/01dq60k83", "country_code": "JP", "type": "education", "lineage": ["https://openalex.org/I201537933"]}], "countries": ["JP"], "is_corresponding": false, "raw_author_name": "Kobayashi, Kosei", "raw_affiliation_strings": ["Graduate School of Life Sciences , Tohoku University , Katahira 2-1-1 , Aoba-ku , Sendai , Miyagi 980-8577 , Japan"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Graduate School of Life Sciences , Tohoku University , Katahira 2-1-1 , Aoba-ku , Sendai , Miyagi 980-8577 , Japan", "institution_ids": ["https://openalex.org/I201537933"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Matsuzaki, Kosuke", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I201537933", "display_name": "Tohoku University", "ror": "https://ror.org/01dq60k83", "country_code": "JP", "type": "education", "lineage": ["https://openalex.org/I201537933"]}], "countries": ["JP"], "is_corresponding": false, "raw_author_name": "Matsuzaki, Kosuke", "raw_affiliation_strings": ["Graduate School of Information Sciences , Tohoku University , 6-6-05 Aramaki Aza Aoba , Aoba-ku , Sendai , 980-8579 , Japan"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Graduate School of Information Sciences , Tohoku University , 6-6-05 Aramaki Aza Aoba , Aoba-ku , Sendai , 980-8579 , Japan", "institution_ids": ["https://openalex.org/I201537933"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Taniguchi, Masaya", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4210126580", "display_name": "RIKEN Center for Advanced Intelligence Project", "ror": "https://ror.org/03ckxwf91", "country_code": "JP", "type": "facility", "lineage": ["https://openalex.org/I4210110652", "https://openalex.org/I4210126580"]}], "countries": ["JP"], "is_corresponding": false, "raw_author_name": "Taniguchi, Masaya", "raw_affiliation_strings": ["RIKEN Center for Advanced Intelligence Project , RIKEN , 1-4-1 Nihonbashi , Chuo-ku , Tokyo , 103-0027 , Japan"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "RIKEN Center for Advanced Intelligence Project , RIKEN , 1-4-1 Nihonbashi , Chuo-ku , Tokyo , 103-0027 , Japan", "institution_ids": ["https://openalex.org/I4210126580"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101067919", "display_name": "Keisuke Sakaguchi", "orcid": "https://orcid.org/0000-0002-3809-1732", "observed_orcids": ["https://orcid.org/0000-0002-3809-1732"]}, "institutions": [{"id": "https://openalex.org/I4210113480", "display_name": "Mohamed bin Zayed University of Artificial Intelligence", "ror": "https://ror.org/0258gkt32", "country_code": "AE", "type": "education", "lineage": ["https://openalex.org/I4210113480"]}], "countries": ["AE"], "is_corresponding": false, "raw_author_name": "Sakaguchi, Keisuke", "raw_affiliation_strings": ["Natural Language Processing Department , Mohamed bin Zayed University of Artificial Intelligence ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Natural Language Processing Department , Mohamed bin Zayed University of Artificial Intelligence ,", "institution_ids": ["https://openalex.org/I4210113480"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101815181", "display_name": "Kentaro Inui", "orcid": "https://orcid.org/0000-0001-6510-604X", "observed_orcids": ["https://orcid.org/0000-0001-6510-604X"]}, "institutions": [{"id": "https://openalex.org/I201537933", "display_name": "Tohoku University", "ror": "https://ror.org/01dq60k83", "country_code": "JP", "type": "education", "lineage": ["https://openalex.org/I201537933"]}], "countries": ["JP"], "is_corresponding": false, "raw_author_name": "Inui, Kentaro", "raw_affiliation_strings": ["Center for Language AI Research , Tohoku University , 6-6-05 Aramaki Aza Aoba , Aoba-ku , Sendai , 980-8579 , Japan"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Center for Language AI Research , Tohoku University , 6-6-05 Aramaki Aza Aoba , Aoba-ku , Sendai , 980-8579 , Japan", "institution_ids": ["https://openalex.org/I201537933"]}]}, {"author_position": "last", "author": {"id": null, "display_name": "Abe, Kentaro", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I201537933", "display_name": "Tohoku University", "ror": "https://ror.org/01dq60k83", "country_code": "JP", "type": "education", "lineage": ["https://openalex.org/I201537933"]}, {"id": "https://openalex.org/I4210113480", "display_name": "Mohamed bin Zayed University of Artificial Intelligence", "ror": "https://ror.org/0258gkt32", "country_code": "AE", "type": "education", "lineage": ["https://openalex.org/I4210113480"]}], "countries": ["AE", "JP"], "is_corresponding": false, "raw_author_name": "Abe, Kentaro", "raw_affiliation_strings": ["Division for the Establishment of Frontier Sciences , Tohoku University , Katahira 2-1-1 , Aoba-ku , Sendai , Miyagi 980-8577 , Japan", "Natural Language Processing Department , Mohamed bin Zayed University of Artificial Intelligence ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Division for the Establishment of Frontier Sciences , Tohoku University , Katahira 2-1-1 , Aoba-ku , Sendai , Miyagi 980-8577 , Japan", "institution_ids": ["https://openalex.org/I201537933"]}, {"raw_affiliation_string": "Natural Language Processing Department , Mohamed bin Zayed University of Artificial Intelligence ,", "institution_ids": ["https://openalex.org/I4210113480"]}]}], "abstract_inverted_index": {"The": [0], "long-range": [1, 43, 116], "dependencies": [2, 21, 44, 117], "among": [3, 45], "the": [4, 24, 60, 65, 126, 145, 153, 159, 169, 180, 191], "tokens,": [5, 46], "which": [6, 73, 100], "originate": [7], "from": [8], "hierarchical": [9], "structures,": [10], "are": [11, 74], "a": [12, 31, 48, 90, 95, 175], "defining": [13], "hallmark": [14], "of": [15, 27, 33, 67, 98, 128, 147, 162, 171, 183], "human": [16], "language.": [17], "However,": [18], "whether": [19], "similar": [20], "exist": [22], "within": [23, 118], "sequential": [25], "vocalization": [26], "non-human": [28, 184], "animals": [29], "remains": [30], "topic": [32], "investigation.": [34], "Transformer": [35, 61], "architectures,": [36], "known": [37], "for": [38, 51, 178], "their": [39, 77], "ability": [40], "to": [41, 63], "model": [42, 92], "provide": [47], "powerful": [49], "tool": [50], "investigating": [52], "this": [53, 56, 85, 106], "phenomenon.": [54], "In": [55], "study,": [57], "we": [58, 87], "employed": [59], "architecture": [62, 103], "analyze": [64], "songs": [66], "Bengalese": [68], "finch": [69], "(Lonchura": [70], "striata": [71], "domestica),": [72], "characterized": [75], "by": [76], "highly": [78], "variable": [79], "and": [80, 130, 140, 197], "complex": [81], "syllable": [82], "sequences.": [83, 120], "To": [84], "end,": [86], "developed": [88], "FinchGPT,": [89], "Transformer-based": [91], "trained": [93], "on": [94, 133, 190], "textualized": [96], "corpus": [97], "birdsongs,": [99], "outperformed": [101], "other": [102], "models": [104, 165], "in": [105, 167], "domain.": [107], "Attention": [108], "weight": [109], "analysis": [110], "revealed": [111], "that": [112], "FinchGPT": [113], "effectively": [114], "captures": [115], "syllables": [119], "Furthermore,": [121], "reverse": [122], "engineering": [123], "approaches": [124], "demonstrated": [125], "impact": [127], "computational": [129, 192], "biological": [131, 195], "manipulations": [132], "its": [134], "performance:": [135], "restricting": [136], "FinchGPT's": [137], "attention": [138], "span": [139], "disrupting": [141], "birdsong": [142], "syntax": [143], "through": [144], "ablation": [146], "specific": [148], "brain": [149], "nuclei": [150], "markedly": [151], "influenced": [152], "model's": [154], "outputs.": [155], "Our": [156], "study": [157], "highlights": [158], "transformative": [160], "potential": [161], "large": [163], "language": [164], "(LLMs)": [166], "deciphering": [168], "complexities": [170], "animal": [172], "vocalizations,": [173], "offering": [174], "novel": [176], "framework": [177], "exploring": [179], "structural": [181], "properties": [182], "communication": [185], "systems": [186], "while": [187], "shedding": [188], "light": [189], "distinctions": [193], "between": [194], "brains": [196], "artificial": [198], "neural": [199], "networks.": [200]}, "primary_topic": {"id": "https://openalex.org/T11665", "display_name": "Animal Vocal Communication and Behavior", "score": 0.9970999956130981, "subfield": {"id": "https://openalex.org/subfields/1309", "display_name": "Developmental Biology"}, "field": {"id": "https://openalex.org/fields/13", "display_name": "Biochemistry, Genetics and Molecular Biology"}, "domain": {"id": "https://openalex.org/domains/1", "display_name": "Life Sciences"}}, "topics": [{"id": "https://openalex.org/T11665", "display_name": "Animal Vocal Communication and Behavior", "score": 0.9970999956130981, "subfield": {"id": "https://openalex.org/subfields/1309", "display_name": "Developmental Biology"}, "field": {"id": "https://openalex.org/fields/13", "display_name": "Biochemistry, Genetics and Molecular Biology"}, "domain": {"id": "https://openalex.org/domains/1", "display_name": "Life Sciences"}}, {"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.0010000000474974513, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11309", "display_name": "Music and Audio Processing", "score": 0.0005000000237487257, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.00344", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.00344", "pdf_url": "https://arxiv.org/pdf/2502.00344", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-nd", "license_id": "https://openalex.org/licenses/cc-by-nc-nd", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4415318248", "doi": "https://doi.org/10.48550/arxiv.2510.07293", "title": "AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMs", "publication_year": 2025, "publication_date": "2025-10-08", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "He, Peize", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "He, Peize", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102635197", "display_name": "Zichen Wen", "orcid": "https://orcid.org/0009-0002-6157-5898", "observed_orcids": ["https://orcid.org/0009-0002-6157-5898"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}, {"id": "https://openalex.org/I4391012619", "display_name": "Shanghai Artificial Intelligence Laboratory", "ror": "https://ror.org/03wkvpx79", "country_code": "CN", "type": "facility", "lineage": ["https://openalex.org/I4391012619"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wen, Zichen", "raw_affiliation_strings": ["Shanghai AI Laboratory", "Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai AI Laboratory", "institution_ids": ["https://openalex.org/I4391012619"]}, {"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100383178", "display_name": "Yubo Wang", "orcid": "https://orcid.org/0000-0001-5766-3751", "observed_orcids": ["https://orcid.org/0000-0001-5766-3751", "https://orcid.org/0009-0001-2313-6299"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Yubo", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100375974", "display_name": "Yuxuan Wang", "orcid": "https://orcid.org/0000-0001-6476-9589", "observed_orcids": ["https://orcid.org/0000-0001-6476-9589", "https://orcid.org/0009-0008-2055-7785", "https://orcid.org/0009-0006-6417-5207"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Yuxuan", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103231035", "display_name": "Xiaoqian Liu", "orcid": "https://orcid.org/0000-0002-2814-3404", "observed_orcids": ["https://orcid.org/0000-0002-2814-3404"]}, "institutions": [{"id": "https://openalex.org/I12912129", "display_name": "Northeastern University", "ror": "https://ror.org/04t5xt781", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I12912129"]}, {"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN", "US"], "is_corresponding": false, "raw_author_name": "Liu, Xiaoqian", "raw_affiliation_strings": ["Northeastern University", "Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Northeastern University", "institution_ids": ["https://openalex.org/I12912129"]}, {"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045031730", "display_name": "Jiajie Huang", "orcid": "https://orcid.org/0000-0002-7788-249X", "observed_orcids": ["https://orcid.org/0000-0002-7788-249X"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Huang, Jiajie", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Lei, Zehui", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Lei, Zehui", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5104287356", "display_name": "Zhuangcheng Gu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I74973139", "display_name": "Carnegie Mellon University", "ror": "https://ror.org/05x2bcf33", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74973139"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Gu, Zhuangcheng", "raw_affiliation_strings": ["Carnegie Mellon University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Carnegie Mellon University", "institution_ids": ["https://openalex.org/I74973139"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Jin, Xiangqi", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Jin, Xiangqi", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Yang, Jiabing", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4210165038", "display_name": "University of Chinese Academy of Sciences", "ror": "https://ror.org/05qbk4x57", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I19820366", "https://openalex.org/I4210165038"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yang, Jiabing", "raw_affiliation_strings": ["University of Chinese Academy of Sciences"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Chinese Academy of Sciences", "institution_ids": ["https://openalex.org/I4210165038"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Li, Kai", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Li, Kai", "raw_affiliation_strings": ["Tsinghua University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Liu, Zhifei", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Liu, Zhifei", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100613916", "display_name": "Weijia Li", "orcid": "https://orcid.org/0000-0003-0876-9773", "observed_orcids": ["https://orcid.org/0000-0003-0876-9773"]}, "institutions": [{"id": "https://openalex.org/I157773358", "display_name": "Sun Yat-sen University", "ror": "https://ror.org/0064kty71", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I157773358"]}, {"id": "https://openalex.org/I4391012619", "display_name": "Shanghai Artificial Intelligence Laboratory", "ror": "https://ror.org/03wkvpx79", "country_code": "CN", "type": "facility", "lineage": ["https://openalex.org/I4391012619"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Li, Weijia", "raw_affiliation_strings": ["Shanghai AI Laboratory", "Sun Yat-sen University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai AI Laboratory", "institution_ids": ["https://openalex.org/I4391012619"]}, {"raw_affiliation_string": "Sun Yat-sen University", "institution_ids": ["https://openalex.org/I157773358"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5010758235", "display_name": "Cunxiang Wang", "orcid": "https://orcid.org/0000-0002-3023-8082", "observed_orcids": ["https://orcid.org/0000-0002-3023-8082"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Cunxiang", "raw_affiliation_strings": ["Tsinghua University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101615091", "display_name": "Conghui He", "orcid": "https://orcid.org/0000-0001-8697-695X", "observed_orcids": ["https://orcid.org/0000-0001-8697-695X"]}, "institutions": [{"id": "https://openalex.org/I4391012619", "display_name": "Shanghai Artificial Intelligence Laboratory", "ror": "https://ror.org/03wkvpx79", "country_code": "CN", "type": "facility", "lineage": ["https://openalex.org/I4391012619"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "He, Conghui", "raw_affiliation_strings": ["Shanghai AI Laboratory"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai AI Laboratory", "institution_ids": ["https://openalex.org/I4391012619"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5100689115", "display_name": "Linfeng Zhang", "orcid": "https://orcid.org/0000-0002-4239-7953", "observed_orcids": ["https://orcid.org/0000-0002-4239-7953"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhang, Linfeng", "raw_affiliation_strings": ["Shanghai Jiao Tong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiao Tong University", "institution_ids": ["https://openalex.org/I183067930"]}]}], "abstract_inverted_index": {"Processing": [0], "long-form": [1, 66], "audio": [2, 30, 80, 99, 127, 171, 180, 187], "is": [3], "a": [4, 55, 70], "major": [5], "challenge": [6], "for": [7, 158], "Large": [8], "Audio": [9], "Language": [10], "models": [11, 14, 42, 182], "(LALMs).": [12], "These": [13], "struggle": [15], "with": [16, 24, 82], "the": [17, 137, 156, 170], "quadratic": [18], "cost": [19], "of": [20, 73, 95, 139, 184], "attention": [21], "($O(N^2)$)": [22], "and": [23, 38, 62, 108, 110, 121, 135, 142, 154, 162, 172], "modeling": [25], "long-range": [26], "temporal": [27, 160], "dependencies.": [28], "Existing": [29], "benchmarks": [31], "are": [32], "built": [33, 75], "mostly": [34], "from": [35, 85], "short": [36], "clips": [37], "do": [39], "not": [40], "evaluate": [41, 59, 118], "in": [43], "realistic": [44], "long": [45], "context": [46], "settings.": [47], "To": [48], "address": [49], "this": [50], "gap,": [51], "we": [52], "introduce": [53], "AudioMarathon,": [54], "benchmark": [56], "designed": [57], "to": [58, 87, 92, 97, 176], "both": [60], "understanding": [61, 181], "inference": [63], "efficiency": [64], "on": [65], "audio.": [67], "AudioMarathon": [68, 167], "provides": [69], "diverse": [71], "set": [72], "tasks": [74], "upon": [76], "three": [77], "pillars:": [78], "long-context": [79], "inputs": [81], "durations": [83], "ranging": [84], "90.0": [86], "300.0": [88], "seconds,": [89], "which": [90], "correspond": [91], "encoded": [93], "sequences": [94], "2,250": [96], "7,500": [98], "tokens,": [100], "respectively,": [101], "full": [102], "domain": [103], "coverage": [104], "across": [105, 151], "speech,": [106], "sound,": [107], "music,": [109], "complex": [111, 186], "reasoning": [112, 161], "that": [113], "requires": [114], "multi-hop": [115], "inference.": [116], "We": [117, 130, 165], "state-of-the-art": [119], "LALMs": [120, 153], "observe": [122], "clear": [123], "performance": [124], "drops": [125], "as": [126], "length": [128], "grows.": [129], "also": [131], "study": [132], "acceleration": [133], "techniques": [134], "analyze": [136], "trade-offs": [138], "token": [140], "pruning": [141], "KV": [143], "cache": [144], "eviction.": [145], "The": [146], "results": [147], "show": [148], "large": [149], "gaps": [150], "current": [152], "highlight": [155], "need": [157], "better": [159], "memory-efficient": [163], "architectures.": [164], "believe": [166], "will": [168], "drive": [169], "multimodal": [173], "research": [174], "community": [175], "develop": [177], "more": [178], "advanced": [179], "capable": [183], "solving": [185], "tasks.": [188]}, "primary_topic": {"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.7153000235557556, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.7153000235557556, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11309", "display_name": "Music and Audio Processing", "score": 0.21130000054836273, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10860", "display_name": "Speech and Audio Processing", "score": 0.03280000016093254, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.07293", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2510.07293", "pdf_url": "https://arxiv.org/pdf/2510.07293", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "public-domain", "license_id": "https://openalex.org/licenses/public-domain", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4412643919", "doi": "https://doi.org/10.1016/j.jpha.2025.101410", "title": "HyPepTox-Fuse: An interpretable hybrid framework for accurate peptide toxicity prediction fusing protein language model-based embeddings with conventional descriptors", "publication_year": 2025, "publication_date": "2025-07-24", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Duong Thanh Tran", "orcid": "https://orcid.org/0009-0003-5096-8167", "observed_orcids": ["https://orcid.org/0009-0003-5096-8167"]}, "institutions": [{"id": "https://openalex.org/I848706", "display_name": "Sungkyunkwan University", "ror": "https://ror.org/04q78tk20", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I848706"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Duong Thanh Tran", "raw_affiliation_strings": ["Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea"], "raw_orcid": "https://orcid.org/0009-0003-5096-8167", "affiliations": [{"raw_affiliation_string": "Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea", "institution_ids": ["https://openalex.org/I848706"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5091402542", "display_name": "Nhat Truong Pham", "orcid": "https://orcid.org/0000-0002-8086-6722", "observed_orcids": ["https://orcid.org/0000-0002-8086-6722"]}, "institutions": [{"id": "https://openalex.org/I848706", "display_name": "Sungkyunkwan University", "ror": "https://ror.org/04q78tk20", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I848706"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Nhat Truong Pham", "raw_affiliation_strings": ["Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea"], "raw_orcid": "https://orcid.org/0000-0002-8086-6722", "affiliations": [{"raw_affiliation_string": "Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea", "institution_ids": ["https://openalex.org/I848706"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103040369", "display_name": "Nguyen Doan Hieu Nguyen", "orcid": "https://orcid.org/0009-0003-9259-8600", "observed_orcids": ["https://orcid.org/0009-0003-9259-8600"]}, "institutions": [{"id": "https://openalex.org/I848706", "display_name": "Sungkyunkwan University", "ror": "https://ror.org/04q78tk20", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I848706"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Nguyen Doan Hieu Nguyen", "raw_affiliation_strings": ["Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea", "institution_ids": ["https://openalex.org/I848706"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5033232360", "display_name": "Leyi Wei", "orcid": "https://orcid.org/0000-0003-1444-190X", "observed_orcids": ["https://orcid.org/0000-0003-1444-190X"]}, "institutions": [{"id": "https://openalex.org/I49835588", "display_name": "Macao Polytechnic University", "ror": "https://ror.org/02sf5td35", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I49835588"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Leyi Wei", "raw_affiliation_strings": ["Centre for Artificial Intelligence Driven Drug Discovery, Faculty of Applied Science, Macao Polytechnic University, Macao SAR, 999078, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Centre for Artificial Intelligence Driven Drug Discovery, Faculty of Applied Science, Macao Polytechnic University, Macao SAR, 999078, China", "institution_ids": ["https://openalex.org/I49835588"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5040180866", "display_name": "Balachandran Manavalan", "orcid": "https://orcid.org/0000-0003-0697-9419", "observed_orcids": ["https://orcid.org/0000-0003-0697-9419"]}, "institutions": [{"id": "https://openalex.org/I848706", "display_name": "Sungkyunkwan University", "ror": "https://ror.org/04q78tk20", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I848706"]}], "countries": ["KR"], "is_corresponding": true, "raw_author_name": "Balachandran Manavalan", "raw_affiliation_strings": ["Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea"], "raw_orcid": "https://orcid.org/0000-0003-0697-9419", "affiliations": [{"raw_affiliation_string": "Department of Integrative Biotechnology, College of Biotechnology and Bioengineering, Sungkyunkwan University, Suwon, 16419, Republic of Korea", "institution_ids": ["https://openalex.org/I848706"]}]}], "abstract_inverted_index": {"Peptide-based": [0], "therapeutics": [1], "hold": [2], "great": [3], "promise": [4], "for": [5, 29, 138, 188], "the": [6, 150, 164, 173], "treatment": [7], "of": [8, 24, 154], "various": [9], "diseases;": [10], "however,": [11], "their": [12], "clinical": [13], "application": [14], "is": [15, 27, 167, 176], "often": [16, 61], "hindered": [17], "by": [18, 97], "toxicity": [19, 26, 140, 191], "challenges.": [20], "The": [21, 181], "accurate": [22], "prediction": [23, 120], "peptide": [25, 95, 139, 190], "crucial": [28], "designing": [30], "safe": [31], "peptide-based": [32], "therapeutics.": [33], "While": [34], "traditional": [35], "experimental": [36], "approaches": [37], "are": [38], "time-consuming": [39], "and": [40, 51, 65, 104, 111, 129, 135, 152, 172, 192], "expensive,": [41], "computational": [42], "methods": [43, 60, 126], "have": [44], "emerged": [45], "as": [46], "viable": [47], "alternatives,": [48], "including": [49], "similarity-based": [50], "machine": [52], "learning": [53, 55], "(ML)-/deep": [54], "(DL)-based": [56], "methods.": [57], "However,": [58], "existing": [59], "struggle": [62], "with": [63, 84], "robustness": [64, 151], "generalizability.": [66], "To": [67], "address": [68], "these": [69], "challenges,": [70], "we": [71, 143], "propose": [72], "HyPepTox-Fuse,": [73, 155], "a": [74, 99, 133, 145], "novel": [75], "framework": [76, 123], "that": [77], "fuses": [78], "protein": [79], "language": [80], "model": [81, 161], "(PLM)-based": [82], "embeddings": [83, 91], "conventional": [85, 116], "descriptors.": [86], "HyPepTox-Fuse": [87, 165], "integrates": [88], "ensemble": [89], "PLM-based": [90], "to": [92, 148], "achieve": [93], "richer": [94], "representations": [96], "leveraging": [98], "cross-modal": [100], "multi-head": [101], "attention": [102], "mechanism": [103], "Transformer": [105], "architecture.": [106], "A": [107], "robust": [108], "feature": [109], "ranking": [110], "selection": [112], "pipeline": [113], "further": [114], "refines": [115], "descriptors,": [117], "thus": [118, 183], "enhancing": [119, 160], "performance.": [121, 162], "Our": [122], "outperforms": [124], "state-of-the-art": [125], "in": [127, 159], "cross-validation": [128], "independent": [130], "evaluations,": [131], "offering": [132], "scalable": [134], "reliable": [136], "tool": [137], "prediction.": [141], "Moreover,": [142], "conducted": [144], "case": [146], "study": [147, 182], "validate": [149], "generalizability": [153], "highlighting": [156], "its": [157], "effectiveness": [158], "Furthermore,": [163], "server": [166], "freely": [168], "accessible": [169], "at": [170, 179], "https://balalab-skku.org/HyPepTox-Fuse/": [171], "source": [174], "code": [175], "publicly": [177], "available": [178, 197], "https://github.com/cbbl-skku-org/HyPepTox-Fuse/.": [180], "presents": [184], "an": [185], "intuitive": [186], "platform": [187], "predicting": [189], "supports": [193], "reproducibility": [194], "through": [195], "openly": [196], "datasets.": [198]}, "primary_topic": {"id": "https://openalex.org/T10211", "display_name": "Computational Drug Discovery Methods", "score": 0.8945000171661377, "subfield": {"id": "https://openalex.org/subfields/1703", "display_name": "Computational Theory and Mathematics"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10211", "display_name": "Computational Drug Discovery Methods", "score": 0.8945000171661377, "subfield": {"id": "https://openalex.org/subfields/1703", "display_name": "Computational Theory and Mathematics"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12254", "display_name": "Machine Learning in Bioinformatics", "score": 0.0778999999165535, "subfield": {"id": "https://openalex.org/subfields/1312", "display_name": "Molecular Biology"}, "field": {"id": "https://openalex.org/fields/13", "display_name": "Biochemistry, Genetics and Molecular Biology"}, "domain": {"id": "https://openalex.org/domains/1", "display_name": "Life Sciences"}}, {"id": "https://openalex.org/T12576", "display_name": "vaccines and immunoinformatics approaches", "score": 0.00839999970048666, "subfield": {"id": "https://openalex.org/subfields/1312", "display_name": "Molecular Biology"}, "field": {"id": "https://openalex.org/fields/13", "display_name": "Biochemistry, Genetics and Molecular Biology"}, "domain": {"id": "https://openalex.org/domains/1", "display_name": "Life Sciences"}}], "primary_location": {"id": "doi:10.1016/j.jpha.2025.101410", "is_oa": true, "landing_page_url": "https://doi.org/10.1016/j.jpha.2025.101410", "pdf_url": null, "source": {"id": "https://openalex.org/S2765050281", "display_name": "Journal of Pharmaceutical Analysis", "issn_l": "2095-1779", "issn": ["2095-1779", "2214-0883"], "is_oa": true, "is_in_doaj": true, "is_core": true, "listed_in": ["cwts-core", "doaj", "jufo-1", "ki-jl-1", "norway-1"], "host_organization": "https://openalex.org/P4310320990", "host_organization_name": "Elsevier BV", "host_organization_lineage": ["https://openalex.org/P4310320990"], "host_organization_lineage_names": ["Elsevier BV"], "type": "journal"}, "license": "cc-by-nc-nd", "license_id": "https://openalex.org/licenses/cc-by-nc-nd", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Journal of Pharmaceutical Analysis", "raw_type": "journal-article"}, "cited_by_count": 13, "is_retracted": false}, {"id": "https://openalex.org/W4407571389", "doi": "https://doi.org/10.48550/arxiv.2502.08910", "title": "InfiniteHiP: Extending Language Model Context Up to 3 Million Tokens on a Single GPU", "publication_year": 2025, "publication_date": "2025-02-13", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5001303923", "display_name": "Hee Jun Lee", "orcid": "https://orcid.org/0000-0002-7273-0112", "observed_orcids": ["https://orcid.org/0000-0002-7273-0112"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lee, Heejun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101640227", "display_name": "Geon Park", "orcid": "https://orcid.org/0000-0002-6878-2649", "observed_orcids": ["https://orcid.org/0000-0002-6878-2649"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Park, Geon", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Suh, Jaduk", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Suh, Jaduk", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5070302452", "display_name": "Sung Ju Hwang", "orcid": "https://orcid.org/0000-0002-9675-2324", "observed_orcids": ["https://orcid.org/0000-0002-9675-2324"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hwang, Sung Ju", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"In": [0], "modern": [1], "large": [2], "language": [3], "models": [4], "(LLMs),": [5], "handling": [6], "very": [7], "long": [8], "context": [9, 62, 138, 154], "lengths": [10], "presents": [11], "significant": [12], "challenges": [13], "as": [14], "it": [15], "causes": [16], "slower": [17], "inference": [18, 53], "speeds": [19], "and": [20, 41, 50, 167, 171], "increased": [21], "memory": [22, 102, 108], "costs.": [23], "Additionally,": [24], "most": [25], "existing": [26], "pre-trained": [27], "LLMs": [28], "fail": [29], "to": [30, 76, 87, 100, 119], "generalize": [31], "beyond": [32], "their": [33], "original": [34], "training": [35], "sequence": [36], "lengths.": [37], "To": [38], "enable": [39], "efficient": [40], "practical": [42, 51], "long-context": [43], "utilization,": [44], "we": [45, 95], "introduce": [46], "InfiniteHiP,": [47], "a": [48, 65, 111, 124, 150], "novel,": [49], "LLM": [52], "framework": [54, 141, 166], "that": [55], "accelerates": [56], "processing": [57, 116], "by": [58, 79], "dynamically": [59], "eliminating": [60], "irrelevant": [61], "tokens": [63, 122], "through": [64, 173], "modular": [66], "hierarchical": [67], "token": [68, 153], "pruning": [69], "algorithm.": [70], "Our": [71, 140], "method": [72, 162], "also": [73], "allows": [74], "generalization": [75], "longer": [77], "sequences": [78], "selectively": [80], "applying": [81], "various": [82], "RoPE": [83], "adjustment": [84], "methods": [85], "according": [86], "the": [88, 97, 115, 164], "internal": [89], "attention": [90, 147], "patterns": [91], "within": [92], "LLMs.": [93], "Furthermore,": [94], "offload": [96], "key-value": [98], "cache": [99], "host": [101], "during": [103], "inference,": [104], "significantly": [105], "reducing": [106], "GPU": [107, 128], "pressure.": [109], "As": [110], "result,": [112], "InfiniteHiP": [113], "enables": [114], "of": [117, 137], "up": [118], "3": [120], "million": [121, 152], "on": [123], "single": [125], "L40s": [126], "48GB": [127], "--": [129, 132], "3x": [130], "larger": [131], "without": [133, 155], "any": [134], "permanent": [135], "loss": [136], "information.": [139], "achieves": [142], "an": [143], "18.95x": [144], "speedup": [145], "in": [146, 163], "decoding": [148], "for": [149], "1": [151], "requiring": [156], "additional": [157], "training.": [158], "We": [159], "implement": [160], "our": [161], "SGLang": [165], "demonstrate": [168], "its": [169], "effectiveness": [170], "practicality": [172], "extensive": [174], "evaluations.": [175]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5449000000953674, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5449000000953674, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.4077000021934509, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.011900000274181366, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.08910", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.08910", "pdf_url": "https://arxiv.org/pdf/2502.08910", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416125299", "doi": "https://doi.org/10.48550/arxiv.2504.04150", "title": "Reasoning on Multiple Needles In A Haystack", "publication_year": 2025, "publication_date": "2025-04-05", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5120394271", "display_name": "Yidong Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Wang, Yidong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0], "Needle": [1], "In": [2, 42, 97], "A": [3, 43], "Haystack": [4, 44], "(NIAH)": [5], "task": [6], "has": [7], "been": [8], "widely": [9], "used": [10], "to": [11, 70], "evaluate": [12], "the": [13, 39, 72, 89, 102, 121, 124, 127, 131, 141, 163, 171, 177], "long-context": [14], "question-answering": [15], "capabilities": [16], "of": [17, 54, 74, 126, 179], "Large": [18], "Language": [19], "Models": [20], "(LLMs).": [21], "However,": [22], "its": [23, 29], "reliance": [24], "on": [25, 136, 190], "simple": [26], "retrieval": [27, 145], "limits": [28], "effectiveness.": [30], "To": [31], "address": [32, 71], "this": [33, 64, 98, 137, 180], "limitation,": [34], "recent": [35], "studies": [36], "have": [37], "introduced": [38], "Multiple": [40], "Needles": [41], "Reasoning": [45], "(MNIAH-R)": [46], "task,": [47], "which": [48, 168], "incorporates": [49], "supporting": [50], "documents": [51], "(Multiple": [52], "needles)": [53], "multi-hop": [55], "reasoning": [56, 147, 185], "tasks": [57], "into": [58, 144], "a": [59, 151, 160], "distracting": [60], "context": [61, 94], "(Haystack}).": [62], "Despite": [63], "advancement,": [65], "existing": [66], "approaches": [67], "still": [68], "fail": [69], "issue": [73], "models": [75], "providing": [76], "direct": [77], "answers": [78], "from": [79], "internal": [80], "knowledge,": [81], "and": [82, 111, 146, 149], "they": [83], "do": [84], "not": [85], "explain": [86], "or": [87], "mitigate": [88, 170], "decline": [90], "in": [91, 123, 183], "accuracy": [92], "as": [93, 130], "length": [95, 125, 133], "increases.": [96, 134], "paper,": [99], "we": [100, 112, 139, 175], "tackle": [101], "memory-based": [103], "answering": [104], "problem": [105], "by": [106, 120], "filtering": [107], "out": [108], "direct-answer": [109], "questions,": [110], "reveal": [113], "that": [114], "performance": [115, 172, 189], "degradation": [116], "is": [117], "primarily": [118], "driven": [119], "reduction": [122], "thinking": [128, 142, 166], "process": [129, 143], "input": [132], "Building": [135], "insight,": [138], "decompose": [140], "stages": [148], "introduce": [150], "reflection": [152], "mechanism": [153], "for": [154], "multi-round": [155], "extension.": [156], "We": [157], "also": [158], "train": [159], "model": [161], "using": [162], "generated": [164], "iterative": [165], "process,": [167], "helps": [169], "degradation.": [173], "Furthermore,": [174], "demonstrate": [176], "application": [178], "retrieval-reflection": [181], "capability": [182], "mathematical": [184], "scenarios,": [186], "improving": [187], "GPT-4o's": [188], "AIME2024.": [191]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9301999807357788, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9301999807357788, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.028699999675154686, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.021299999207258224, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2504.04150", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2504.04150", "pdf_url": "https://arxiv.org/pdf/2504.04150", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4411593307", "doi": "https://doi.org/10.1016/j.osn.2025.100812", "title": "Revolutionizing optical networks: The integration and impact of large language models", "publication_year": 2025, "publication_date": "2025-06-24", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5093793521", "display_name": "Sergio Cruzes", "orcid": "https://orcid.org/0009-0008-8955-7628", "observed_orcids": ["https://orcid.org/0009-0008-8955-7628"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Sergio Cruzes", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0008-8955-7628", "affiliations": []}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T10847", "display_name": "Advanced Optical Network Technologies", "score": 0.8246999979019165, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10847", "display_name": "Advanced Optical Network Technologies", "score": 0.8246999979019165, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10714", "display_name": "Software-Defined Networks and 5G", "score": 0.16410000622272491, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10232", "display_name": "Optical Network Technologies", "score": 0.003800000064074993, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1016/j.osn.2025.100812", "is_oa": false, "landing_page_url": "https://doi.org/10.1016/j.osn.2025.100812", "pdf_url": null, "source": {"id": "https://openalex.org/S48659482", "display_name": "Optical Switching and Networking", "issn_l": "1573-4277", "issn": ["1573-4277", "1872-9770"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310320990", "host_organization_name": "Elsevier BV", "host_organization_lineage": ["https://openalex.org/P4310320990"], "host_organization_lineage_names": ["Elsevier BV"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Optical Switching and Networking", "raw_type": "journal-article"}, "cited_by_count": 8, "is_retracted": false}, {"id": "https://openalex.org/W4414851187", "doi": "https://doi.org/10.48550/arxiv.2502.17807", "title": "DocPuzzle: A Process-Aware Benchmark for Evaluating Realistic Long-Context Reasoning Capabilities", "publication_year": 2025, "publication_date": "2025-02-25", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5085025479", "display_name": "Tianyi Zhuang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhuang, Tianyi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Kuang, Chuqiao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kuang, Chuqiao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100373874", "display_name": "Xiaoguang Li", "orcid": "https://orcid.org/0000-0003-4580-4865", "observed_orcids": ["https://orcid.org/0000-0003-4580-4865"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Xiaoguang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102573328", "display_name": "Yihua Teng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Teng, Yihua", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5112882104", "display_name": "Jihao Wu", "orcid": "https://orcid.org/0000-0003-3235-0446", "observed_orcids": ["https://orcid.org/0000-0003-3235-0446"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Jihao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5115592503", "display_name": "Yasheng Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yasheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5046228314", "display_name": "Lifeng Shang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Shang, Lifeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0], "present": [1], "DocPuzzle,": [2], "a": [3, 40], "rigorously": [4], "constructed": [5], "benchmark": [6, 18], "for": [7, 62], "evaluating": [8], "long-context": [9], "reasoning": [10, 26, 64, 75, 92, 109], "capabilities": [11, 110], "in": [12, 66], "large": [13], "language": [14], "models": [15, 76, 86, 93], "(LLMs).": [16], "This": [17], "comprises": [19], "100": [20], "expert-level": [21], "QA": [22], "problems": [23], "requiring": [24], "multi-step": [25], "over": [27], "long": [28], "real-world": [29], "documents.": [30], "To": [31], "ensure": [32], "the": [33, 99, 106], "task": [34], "quality": [35], "and": [36, 79], "complexity,": [37], "we": [38], "implement": [39], "human-AI": [41], "collaborative": [42], "annotation-validation": [43], "pipeline.": [44], "DocPuzzle": [45], "introduces": [46], "an": [47], "innovative": [48], "evaluation": [49, 69], "framework": [50], "that": [51], "mitigates": [52], "guessing": [53], "bias": [54], "through": [55], "checklist-guided": [56], "process": [57], "analysis,": [58], "establishing": [59], "new": [60], "standards": [61], "assessing": [63], "capacities": [65], "LLMs.": [67], "Our": [68], "results": [70], "show": [71], "that:": [72], "1)Advanced": [73], "slow-thinking": [74], "like": [77, 87, 94], "o1-preview(69.7%)": [78], "DeepSeek-R1(66.3%)": [80], "significantly": [81], "outperform": [82], "best": [83], "general": [84], "instruct": [85], "Claude": [88], "3.5": [89], "Sonnet(57.7%);": [90], "2)Distilled": [91], "DeepSeek-R1-Distill-Qwen-32B(41.3%)": [95], "falls": [96], "far": [97], "behind": [98], "teacher": [100], "model,": [101], "suggesting": [102], "challenges": [103], "to": [104], "maintain": [105], "generalization": [107], "of": [108], "relying": [111], "solely": [112], "on": [113], "distillation.": [114]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9491999745368958, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9491999745368958, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.021800000220537186, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.01810000091791153, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.17807", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.17807", "pdf_url": "https://arxiv.org/pdf/2502.17807", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416692015", "doi": "https://doi.org/10.1016/j.trc.2025.105456", "title": "Addressing corner cases in autonomous driving: A world model-based approach with mixture of experts and LLMs", "publication_year": 2025, "publication_date": "2025-11-26", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5032304100", "display_name": "Haicheng Liao", "orcid": "https://orcid.org/0000-0002-0207-5087", "observed_orcids": ["https://orcid.org/0000-0002-0207-5087"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Haicheng Liao", "raw_affiliation_strings": ["Department of Computer and Information Science, University of Macau, Macau SAR, China", "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0000-0002-0207-5087", "affiliations": [{"raw_affiliation_string": "Department of Computer and Information Science, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045228584", "display_name": "Bonan Wang", "orcid": "https://orcid.org/0009-0005-8271-0341", "observed_orcids": ["https://orcid.org/0009-0005-8271-0341"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Bonan Wang", "raw_affiliation_strings": ["Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "Department of Computer and Information Science, University of Macau, Macau SAR, China", "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0009-0005-8271-0341", "affiliations": [{"raw_affiliation_string": "Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "Department of Computer and Information Science, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015192790", "display_name": "Jae-Ung Yang", "orcid": "https://orcid.org/0009-0007-7846-463X", "observed_orcids": ["https://orcid.org/0009-0007-7846-463X"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Junxian Yang", "raw_affiliation_strings": ["State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0009-0007-7846-463X", "affiliations": [{"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5106627873", "display_name": "Chengyue Wang", "orcid": "https://orcid.org/0009-0009-7707-1505", "observed_orcids": ["https://orcid.org/0009-0009-7707-1505", "https://orcid.org/0009-0000-6957-5468"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Chengyue Wang", "raw_affiliation_strings": ["Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0009-0009-7707-1505", "affiliations": [{"raw_affiliation_string": "Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhengbing He", "orcid": "https://orcid.org/0000-0001-5716-3853", "observed_orcids": ["https://orcid.org/0000-0001-5716-3853"]}, "institutions": [{"id": "https://openalex.org/I63966007", "display_name": "Massachusetts Institute of Technology", "ror": "https://ror.org/042nb2s44", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I63966007"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Zhengbing He", "raw_affiliation_strings": ["Senseable City Lab, Massachusetts Institute of Technology, Cambridge MA, United States"], "raw_orcid": "https://orcid.org/0000-0001-5716-3853", "affiliations": [{"raw_affiliation_string": "Senseable City Lab, Massachusetts Institute of Technology, Cambridge MA, United States", "institution_ids": ["https://openalex.org/I63966007"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5055081255", "display_name": "Guohui Zhang", "orcid": "https://orcid.org/0000-0001-5194-9222", "observed_orcids": ["https://orcid.org/0000-0001-5194-9222"]}, "institutions": [{"id": "https://openalex.org/I117965899", "display_name": "University of Hawai\u02bbi at M\u0101noa", "ror": "https://ror.org/01wspgy28", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I117965899"]}, {"id": "https://openalex.org/I1331384533", "display_name": "University of Hawaii System", "ror": "https://ror.org/03tzaeb71", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1331384533"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Guohui Zhang", "raw_affiliation_strings": ["Department of Civil and Environmental Engineering, University of Hawaii, Honolulu HI, United States"], "raw_orcid": "https://orcid.org/0000-0001-5194-9222", "affiliations": [{"raw_affiliation_string": "Department of Civil and Environmental Engineering, University of Hawaii, Honolulu HI, United States", "institution_ids": ["https://openalex.org/I117965899", "https://openalex.org/I1331384533"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5012773300", "display_name": "Chengzhong Xu", "orcid": "https://orcid.org/0000-0001-9480-0356", "observed_orcids": ["https://orcid.org/0000-0001-9480-0356"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": false, "raw_author_name": "Chengzhong Xu", "raw_affiliation_strings": ["Department of Computer and Information Science, University of Macau, Macau SAR, China", "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0000-0001-9480-0356", "affiliations": [{"raw_affiliation_string": "Department of Computer and Information Science, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101552930", "display_name": "Zhenning Li", "orcid": "https://orcid.org/0000-0002-0877-6829", "observed_orcids": ["https://orcid.org/0000-0002-0877-6829"]}, "institutions": [{"id": "https://openalex.org/I204512498", "display_name": "University of Macau", "ror": "https://ror.org/01r4q9n85", "country_code": "MO", "type": "education", "lineage": ["https://openalex.org/I204512498"]}], "countries": ["MO"], "is_corresponding": true, "raw_author_name": "Zhenning Li", "raw_affiliation_strings": ["Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "Department of Computer and Information Science, University of Macau, Macau SAR, China", "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China"], "raw_orcid": "https://orcid.org/0000-0002-0877-6829", "affiliations": [{"raw_affiliation_string": "Department of Civil and Environmental Engineering, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "Department of Computer and Information Science, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}, {"raw_affiliation_string": "State Key Laboratory of Internet of Things for Smart City, University of Macau, Macau SAR, China", "institution_ids": ["https://openalex.org/I204512498"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T11099", "display_name": "Autonomous Vehicle Technology and Safety", "score": 0.9970999956130981, "subfield": {"id": "https://openalex.org/subfields/2203", "display_name": "Automotive Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11099", "display_name": "Autonomous Vehicle Technology and Safety", "score": 0.9970999956130981, "subfield": {"id": "https://openalex.org/subfields/2203", "display_name": "Automotive Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11344", "display_name": "Traffic Prediction and Management Techniques", "score": 0.0007999999797903001, "subfield": {"id": "https://openalex.org/subfields/2215", "display_name": "Building and Construction"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10370", "display_name": "Traffic and Road Safety", "score": 0.00019999999494757503, "subfield": {"id": "https://openalex.org/subfields/2213", "display_name": "Safety, Risk, Reliability and Quality"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1016/j.trc.2025.105456", "is_oa": false, "landing_page_url": "https://doi.org/10.1016/j.trc.2025.105456", "pdf_url": null, "source": {"id": "https://openalex.org/S64511102", "display_name": "Transportation Research Part C Emerging Technologies", "issn_l": "0968-090X", "issn": ["0968-090X", "1879-2359"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["abdc-a-star", "cwts-core", "jufo-3", "norway-1"], "host_organization": "https://openalex.org/P4310320990", "host_organization_name": "Elsevier BV", "host_organization_lineage": ["https://openalex.org/P4310320990"], "host_organization_lineage_names": ["Elsevier BV"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Transportation Research Part C: Emerging Technologies", "raw_type": "journal-article"}, "cited_by_count": 5, "is_retracted": false}, {"id": "https://openalex.org/W4415596407", "doi": "https://doi.org/10.1016/j.aei.2025.104009", "title": "A MoE-LLM-based multisensor flexible fusion fault diagnosis method for rotating machinery", "publication_year": 2025, "publication_date": "2025-10-27", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5042347644", "display_name": "Tantao Lin", "orcid": "https://orcid.org/0000-0003-2589-2634", "observed_orcids": ["https://orcid.org/0000-0003-2589-2634"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Tantao Lin", "raw_affiliation_strings": ["Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0000-0003-2589-2634", "affiliations": [{"raw_affiliation_string": "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5001668408", "display_name": "Zhijun Ren", "orcid": "https://orcid.org/0000-0001-8976-0396", "observed_orcids": ["https://orcid.org/0000-0001-8976-0396"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Zhijun Ren", "raw_affiliation_strings": ["Frontier Institute of Science and Technology, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0000-0001-8976-0396", "affiliations": [{"raw_affiliation_string": "Frontier Institute of Science and Technology, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}, {"raw_affiliation_string": "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102853240", "display_name": "Kai Huang", "orcid": "https://orcid.org/0009-0007-7974-334X", "observed_orcids": ["https://orcid.org/0009-0007-7974-334X"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Kai Huang", "raw_affiliation_strings": ["Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0009-0007-7974-334X", "affiliations": [{"raw_affiliation_string": "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5090855504", "display_name": "Hamid Reza Karimi", "orcid": "https://orcid.org/0000-0001-7629-3266", "observed_orcids": ["https://orcid.org/0000-0001-7629-3266"]}, "institutions": [{"id": "https://openalex.org/I93860229", "display_name": "Politecnico di Milano", "ror": "https://ror.org/01nffqt88", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I93860229"]}], "countries": ["IT"], "is_corresponding": false, "raw_author_name": "Hamid Reza Karimi", "raw_affiliation_strings": ["Department of Mechanical Engineering, Politecnico di Milano, Milano, 20156, Lombardy, Italy"], "raw_orcid": "https://orcid.org/0000-0001-7629-3266", "affiliations": [{"raw_affiliation_string": "Department of Mechanical Engineering, Politecnico di Milano, Milano, 20156, Lombardy, Italy", "institution_ids": ["https://openalex.org/I93860229"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100732893", "display_name": "Yongsheng Zhu", "orcid": "https://orcid.org/0000-0002-9256-2427", "observed_orcids": ["https://orcid.org/0000-0002-9256-2427"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Yongsheng Zhu", "raw_affiliation_strings": ["Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0000-0002-9256-2427", "affiliations": [{"raw_affiliation_string": "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5079329106", "display_name": "Ke Feng", "orcid": "https://orcid.org/0000-0003-2338-5161", "observed_orcids": ["https://orcid.org/0000-0003-2338-5161"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Ke Feng", "raw_affiliation_strings": ["School of Mechanical Engineering, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0000-0003-2338-5161", "affiliations": [{"raw_affiliation_string": "School of Mechanical Engineering, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5027754907", "display_name": "Jun Hong", "orcid": "https://orcid.org/0000-0001-9556-9619", "observed_orcids": ["https://orcid.org/0000-0001-9556-9619"]}, "institutions": [{"id": "https://openalex.org/I87445476", "display_name": "Xi'an Jiaotong University", "ror": "https://ror.org/017zhmm22", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I87445476"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Jun Hong", "raw_affiliation_strings": ["Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China"], "raw_orcid": "https://orcid.org/0000-0001-9556-9619", "affiliations": [{"raw_affiliation_string": "Key Laboratory of Education Ministry for Modern Design and Rotor-Bearing System, Xi\u2019an Jiaotong University, Xi\u2019an, 710049, Shaanxi, China", "institution_ids": ["https://openalex.org/I87445476"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T10220", "display_name": "Machine Fault Diagnosis Techniques", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10220", "display_name": "Machine Fault Diagnosis Techniques", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10876", "display_name": "Fault Detection and Control Systems", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T13857", "display_name": "Mining and Industrial Processes", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1106", "display_name": "Food Science"}, "field": {"id": "https://openalex.org/fields/11", "display_name": "Agricultural and Biological Sciences"}, "domain": {"id": "https://openalex.org/domains/1", "display_name": "Life Sciences"}}], "primary_location": {"id": "doi:10.1016/j.aei.2025.104009", "is_oa": false, "landing_page_url": "https://doi.org/10.1016/j.aei.2025.104009", "pdf_url": null, "source": {"id": "https://openalex.org/S112141509", "display_name": "Advanced Engineering Informatics", "issn_l": "1474-0346", "issn": ["1474-0346", "1873-5320"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310320990", "host_organization_name": "Elsevier BV", "host_organization_lineage": ["https://openalex.org/P4310320990"], "host_organization_lineage_names": ["Elsevier BV"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Advanced Engineering Informatics", "raw_type": "journal-article"}, "cited_by_count": 10, "is_retracted": false}, {"id": "https://openalex.org/W7161476200", "doi": "https://doi.org/10.61838/rpct.3.3.6", "title": "Deep Learning Analysis of Trauma-Related Emotional Patterns in Couples Facing PTSD", "publication_year": 2025, "publication_date": "2025-01-01", "type": "article", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5136338530", "display_name": "Lawrence Siegel", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I111088046", "display_name": "Boston University", "ror": "https://ror.org/05qwgg493", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I111088046"]}], "countries": ["US"], "is_corresponding": true, "raw_author_name": "Lawrence Siegel", "raw_affiliation_strings": ["The Albert and JessieDanielsen Institute, Boston University, Boston, Massachusetts, USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Albert and JessieDanielsen Institute, Boston University, Boston, Massachusetts, USA", "institution_ids": ["https://openalex.org/I111088046"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5126381648", "display_name": "Przemyslaw Wandzilak", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I60158472", "display_name": "Concordia University", "ror": "https://ror.org/0420zvk78", "country_code": "CA", "type": "education", "lineage": ["https://openalex.org/I60158472"]}], "countries": ["CA"], "is_corresponding": false, "raw_author_name": "Przemyslaw Wandzilak", "raw_affiliation_strings": ["Department of Psychology, Concordia University, Montreal, Canada"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Psychology, Concordia University, Montreal, Canada", "institution_ids": ["https://openalex.org/I60158472"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5136323698", "display_name": "Marina Gonalons-Pons", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I79576946", "display_name": "University of Pennsylvania", "ror": "https://ror.org/00b30xv10", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I79576946"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Marina Gonalons-Pons", "raw_affiliation_strings": ["Department of Psychology, University of Pennsylvania, McNeil 217, 3718 Locust Walk, Philadelphia"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Psychology, University of Pennsylvania, McNeil 217, 3718 Locust Walk, Philadelphia", "institution_ids": ["https://openalex.org/I79576946"]}]}], "abstract_inverted_index": {"This": [0], "study": [1, 28], "aimed": [2], "to": [3, 19], "examine": [4], "and": [5, 64, 73, 80, 85, 106, 119, 133, 146, 159, 215, 226, 247, 267, 285, 293], "classify": [6], "trauma-related": [7, 272], "emotional": [8, 22, 117, 135, 166, 206, 245, 252, 273, 283], "patterns": [9, 118, 208, 288], "in": [10, 42, 54, 162, 221, 250, 276], "couples": [11, 35, 277], "facing": [12, 278], "PTSD": [13, 222], "using": [14, 92, 141], "multimodal": [15, 94, 154, 195], "deep": [16, 261], "learning": [17, 262], "models": [18, 161, 185], "predict": [20], "high-risk": [21, 124, 164, 229], "episodes": [23, 125, 167], "during": [24], "conflict": [25], "interactions.": [26], "The": [27, 153], "employed": [29], "a": [30, 55, 93, 127, 239, 264], "cross-sectional": [31], "design": [32], "involving": [33], "112": [34], "(224": [36], "individuals)": [37], "from": [38], "the": [39, 194], "United": [40], "States": [41], "which": [43], "one": [44], "partner": [45], "met": [46], "diagnostic": [47], "criteria": [48], "for": [49, 270], "PTSD.": [50], "Each": [51], "couple": [52], "participated": [53], "40-minute": [56], "conflict-discussion": [57], "task": [58], "recorded": [59], "through": [60], "synchronized": [61], "audio,": [62], "video,": [63], "physiological": [65, 103, 131, 248], "sensors.": [66], "Linguistic": [67], "transcripts,": [68], "facial": [69], "expressions,": [70], "acoustic": [71, 107], "features,": [72, 102], "autonomic": [74], "indicators": [75], "(electrodermal": [76], "activity,": [77], "heart-rate": [78], "variability,": [79], "peripheral": [81], "temperature)": [82], "were": [83, 90], "extracted": [84], "temporally": [86], "aligned.": [87], "These": [88], "modalities": [89], "analyzed": [91], "transformer": [95, 155], "architecture": [96], "integrating": [97], "text": [98], "embeddings,": [99], "CNN-LSTM": [100], "visual": [101], "time-series": [104], "data,": [105], "spectral": [108], "representations.": [109], "Additional": [110], "analyses": [111], "included": [112], "t-SNE\u2013based": [113], "latent": [114], "clustering": [115, 201], "of": [116, 122, 130, 228], "risk-surface": [120, 236], "modeling": [121], "predicted": [123], "as": [126], "nonlinear": [128, 286], "function": [129], "arousal": [132, 249], "negative": [134, 244], "language.": [136], "Model": [137], "performance": [138], "was": [139], "evaluated": [140], "accuracy,": [142], "precision,": [143], "recall,": [144], "F1-score,": [145], "AUC": [147], "metrics": [148], "with": [149], "ten-fold": [150], "stratified": [151], "cross-validation.": [152], "significantly": [156, 191, 220], "outperformed": [157], "baseline": [158], "unimodal": [160, 184], "predicting": [163, 251], "PTSD-related": [165], "(AUC": [168], "=": [169, 172, 175, 255], ".90;": [170], "F1": [171], ".83;": [173], "accuracy": [174], ".86;": [176], "p": [177, 257], "<": [178, 198, 233, 258], ".001": [179], "vs.": [180], "clinical": [181], "baseline).": [182], "All": [183], "performed": [186], "above": [187], "chance": [188], "but": [189], "remained": [190], "weaker": [192], "than": [193], "approach": [196], "(p": [197], ".01).": [199], "Latent": [200], "revealed": [202], "four": [203], "statistically": [204], "distinct": [205], "interaction": [207, 241], "(hyperaroused": [209], "escalation,": [210], "avoidant": [211], "disengagement,": [212], "mixed": [213], "volatile\u2013repairing,": [214], "numbed": [216], "coexistence),": [217], "each": [218], "differing": [219], "severity,": [223], "relationship": [224], "quality,": [225], "proportion": [227], "segments": [230], "(all": [231], "ps": [232], ".05).": [234], "Nonlinear": [235], "analysis": [237], "demonstrated": [238], "strong": [240], "effect": [242], "between": [243], "language": [246], "risk": [253, 274], "(\u03b2interaction": [254], ".41,": [256], ".001).": [259], "Multimodal": [260], "provides": [263], "highly": [265], "sensitive": [266], "integrative": [268], "method": [269], "identifying": [271], "states": [275], "PTSD,": [279], "revealing": [280], "clinically": [281], "meaningful": [282], "clusters": [284], "escalation": [287], "that": [289], "may": [290], "inform": [291], "assessment": [292], "intervention.": [294]}, "primary_topic": {"id": "https://openalex.org/T10242", "display_name": "Posttraumatic Stress Disorder Research", "score": 0.6401000022888184, "subfield": {"id": "https://openalex.org/subfields/3203", "display_name": "Clinical Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T10242", "display_name": "Posttraumatic Stress Disorder Research", "score": 0.6401000022888184, "subfield": {"id": "https://openalex.org/subfields/3203", "display_name": "Clinical Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T10667", "display_name": "Emotion and Mood Recognition", "score": 0.3109999895095825, "subfield": {"id": "https://openalex.org/subfields/3205", "display_name": "Experimental and Cognitive Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T12488", "display_name": "Mental Health via Writing", "score": 0.004399999976158142, "subfield": {"id": "https://openalex.org/subfields/3207", "display_name": "Social Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "doi:10.61838/rpct.3.3.6", "is_oa": true, "landing_page_url": "https://doi.org/10.61838/rpct.3.3.6", "pdf_url": "https://jrpct.com/index.php/rpct/article/download/48/46", "source": {"id": "https://openalex.org/S7407058014", "display_name": "Research and Practice in Couple Therapy", "issn_l": "3092-7684", "issn": ["3092-7684"], "is_oa": false, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": null, "host_organization_name": null, "host_organization_lineage": [], "host_organization_lineage_names": [], "type": "journal"}, "license": "cc-by-nc", "license_id": "https://openalex.org/licenses/cc-by-nc", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Research and Practice in Couple Therapy", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4409530322", "doi": "https://doi.org/10.21203/rs.3.rs-6277319/v1", "title": "From News to Trends: A Financial Time Series Forecasting Framework with LLM-Driven News Sentiment Analysis and Selective State Spaces", "publication_year": 2025, "publication_date": "2025-04-17", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5040675777", "display_name": "Renjie Wang", "orcid": "https://orcid.org/0009-0005-0669-7878", "observed_orcids": ["https://orcid.org/0009-0005-0669-7878"]}, "institutions": [{"id": "https://openalex.org/I194450716", "display_name": "Jilin University", "ror": "https://ror.org/00js3aw79", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I194450716"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Renjie Wang", "raw_affiliation_strings": ["Jilin University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Jilin University", "institution_ids": ["https://openalex.org/I194450716"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070021804", "display_name": "Minghui Sun", "orcid": "https://orcid.org/0000-0002-1809-8187", "observed_orcids": ["https://orcid.org/0000-0002-1809-8187"]}, "institutions": [{"id": "https://openalex.org/I194450716", "display_name": "Jilin University", "ror": "https://ror.org/00js3aw79", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I194450716"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Minghui Sun", "raw_affiliation_strings": ["Jilin University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Jilin University", "institution_ids": ["https://openalex.org/I194450716"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5100436492", "display_name": "Limin Wang", "orcid": "https://orcid.org/0000-0001-9618-9239", "observed_orcids": ["https://orcid.org/0000-0001-9618-9239"]}, "institutions": [{"id": "https://openalex.org/I194450716", "display_name": "Jilin University", "ror": "https://ror.org/00js3aw79", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I194450716"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Limin Wang", "raw_affiliation_strings": ["Jilin University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Jilin University", "institution_ids": ["https://openalex.org/I194450716"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T11326", "display_name": "Stock Market Forecasting Methods", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1803", "display_name": "Management Science and Operations Research"}, "field": {"id": "https://openalex.org/fields/18", "display_name": "Decision Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T11326", "display_name": "Stock Market Forecasting Methods", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1803", "display_name": "Management Science and Operations Research"}, "field": {"id": "https://openalex.org/fields/18", "display_name": "Decision Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T12205", "display_name": "Time Series Analysis and Forecasting", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10047", "display_name": "Financial Markets and Investment Strategies", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/2003", "display_name": "Finance"}, "field": {"id": "https://openalex.org/fields/20", "display_name": "Economics, Econometrics and Finance"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "doi:10.21203/rs.3.rs-6277319/v1", "is_oa": true, "landing_page_url": "https://doi.org/10.21203/rs.3.rs-6277319/v1", "pdf_url": "https://www.researchsquare.com/article/rs-6277319/latest.pdf", "source": {"id": "https://openalex.org/S4306525896", "display_name": "Research Square", "issn_l": null, "issn": null, "is_oa": false, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": null, "host_organization_name": null, "host_organization_lineage": [], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": "Research Square", "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7109543021", "doi": "https://doi.org/10.5281/zenodo.17835987", "title": "Dynamic Semantic Categorization Through Self-Referential Attention: Why the Transformer Excels at Open-World Intelligence", "publication_year": 2025, "publication_date": "2025-12-06", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Wang, Zhongren", "orcid": "https://orcid.org/0009-0002-5225-091X", "observed_orcids": ["https://orcid.org/0009-0002-5225-091X"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Wang, Zhongren", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0002-5225-091X", "affiliations": []}], "abstract_inverted_index": {"The": [0], "evolution": [1], "of": [2, 17, 32, 78], "artificial": [3], "intelligence\u2014from": [4], "expert": [5], "systems": [6], "to": [7], "large": [8], "language": [9], "models\u2014can": [10], "be": [11], "productively": [12], "viewed": [13], "through": [14], "the": [15, 39, 71, 109], "lens": [16], "dynamic": [18, 80], "semantic": [19], "categorization.": [20], "This": [21], "paper": [22], "argues": [23], "that": [24, 38, 107], "perception,": [25], "reasoning,": [26], "and": [27, 52, 59, 83, 102], "generation": [28], "are": [29], "all": [30], "manifestations": [31], "context-sensitive": [33], "categorization": [34], "processes.": [35], "We": [36, 87], "demonstrate": [37], "Transformer": [40], "architecture,": [41], "with": [42], "its": [43], "query-key-value": [44], "(QKV)": [45], "self-attention": [46], "mechanism,": [47], "provides": [48], "a": [49, 75, 104], "particularly": [50], "efficient": [51], "expressive": [53], "computational": [54], "substrate": [55], "for": [56, 69], "open-world,": [57], "context-dependent,": [58], "self-referential": [60], "category": [61], "construction.": [62], "While": [63], "alternative": [64], "architectures": [65], "possess": [66], "varying": [67], "capabilities": [68], "categorization,": [70], "Transformer's": [72], "design": [73], "enables": [74], "unique": [76], "combination": [77], "parallelizable": [79], "prototype": [81], "formation": [82], "global": [84], "context": [85], "integration.": [86], "ground": [88], "this": [89], "analysis": [90], "in": [91], "Wang's": [92], "(2025)": [93], "\"AI": [94], "=": [95], "Dynamic": [96], "Categorization\"": [97], "framework,": [98], "provide": [99], "theoretical": [100], "analysis,": [101], "present": [103], "minimal": [105], "implementation": [106], "illustrates": [108], "core": [110], "mechanism.": [111]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9413999915122986, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9413999915122986, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.02239999920129776, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.009200000204145908, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.5281/zenodo.17835987", "is_oa": true, "landing_page_url": "https://doi.org/10.5281/zenodo.17835987", "pdf_url": null, "source": {"id": "https://openalex.org/S4306400562", "display_name": "Zenodo (CERN European Organization for Nuclear Research)", "issn_l": null, "issn": null, "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I67311998", "host_organization_name": "European Organization for Nuclear Research", "host_organization_lineage": ["https://openalex.org/I67311998"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": null, "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "Preprint"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415082828", "doi": "https://doi.org/10.48550/arxiv.2503.00634", "title": "Efficiently Editing Mixture-of-Experts Models with Compressed Experts", "publication_year": 2025, "publication_date": "2025-03-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5101137979", "display_name": "Yifei He", "orcid": "https://orcid.org/0000-0003-0713-6376", "observed_orcids": ["https://orcid.org/0000-0003-0713-6376"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "He, Yifei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100657670", "display_name": "Yang Liu", "orcid": "https://orcid.org/0000-0002-1297-676X", "observed_orcids": ["https://orcid.org/0000-0002-1297-676X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Yang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5035634586", "display_name": "Liang Chen", "orcid": "https://orcid.org/0000-0002-5570-5820", "observed_orcids": ["https://orcid.org/0000-0002-5570-5820"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liang, Chen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5030937723", "display_name": "Hany Hassan Awadalla", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Awadalla, Hany Hassan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Mixture-of-Experts": [0], "(MoE)": [1], "models": [2, 12, 67, 148, 187, 196], "have": [3], "become": [4], "a": [5, 17, 32], "key": [6], "approach": [7, 117, 181], "for": [8, 68], "scaling": [9, 193], "large": [10], "language": [11], "efficiently": [13], "by": [14], "activating": [15], "only": [16], "subset": [18], "of": [19, 28, 74, 89, 103, 113, 134, 160, 185], "experts": [20, 30, 35, 41, 51, 80, 91, 122, 128, 156], "during": [21], "training": [22], "and": [23, 76, 151, 174, 191], "inference.": [24], "Typically,": [25], "the": [26, 87, 101, 119], "number": [27, 88], "activated": [29, 50, 90, 127], "presents": [31], "trade-off:": [33], "fewer": [34], "reduce": [36, 86], "computational": [37], "costs,": [38], "while": [39, 92, 123, 141, 167], "more": [40, 169], "improve": [42], "performance.": [43, 95, 144], "Recent": [44], "studies": [45], "reveal": [46], "that": [47, 108, 154], "not": [48], "all": [49], "contribute": [52], "equally": [53], "to": [54, 85, 194], "model": [55, 94], "performance,": [56], "with": [57, 129, 197], "some": [58], "providing": [59], "minimal": [60], "utility,": [61], "particularly": [62], "when": [63], "finetuning": [64], "pretrained": [65], "MoE": [66, 186], "specialized": [69], "downstream": [70], "tasks.": [71], "The": [72, 132], "co-existence": [73], "significant": [75], "redundant": [77], "parameters": [78, 136, 173], "in": [79, 177, 188], "provides": [81], "us": [82], "an": [83], "opportunity": [84], "maintaining": [93], "In": [96], "this": [97], "work,": [98], "we": [99], "propose": [100], "concept": [102], "compressed": [104, 130, 155], "experts,": [105], "lightweight": [106], "modules": [107], "serve": [109], "as": [110], "compact": [111], "representations": [112], "full": [114, 161], "experts.": [115, 131], "Our": [116, 200], "preserves": [118], "most": [120], "important": [121], "replacing": [124], "other": [125], "auxiliary": [126], "reduction": [133], "active": [135, 172], "significantly": [137], "lowers": [138], "inference": [139, 178], "costs": [140], "achieving": [142], "comparable": [143], "Extensive": [145], "experiments": [146], "on": [147], "including": [149], "Phi-MoE": [150], "OLMoE": [152], "demonstrate": [153], "recover": [157], "over": [158], "90%": [159], "expert": [162], "performance": [163], "across": [164], "various": [165], "tasks": [166], "reducing": [168], "than": [170], "30%": [171], "saving": [175], "20%": [176], "costs.": [179], "This": [180], "enables": [182], "efficient": [183], "deployment": [184], "resource-constrained": [189], "settings": [190], "facilitates": [192], "larger": [195], "manageable": [198], "overhead.": [199], "code": [201], "is": [202], "available": [203], "at": [204], "https://github.com/yifei-he/Compressed-Experts.": [205]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.42719998955726624, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.42719998955726624, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.4196999967098236, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.08910000324249268, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.00634", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.00634", "pdf_url": "https://arxiv.org/pdf/2503.00634", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414823914", "doi": "https://doi.org/10.48550/arxiv.2510.02194", "title": "UpSafe$^\\circ$C: Upcycling for Controllable Safety in Large Language Models", "publication_year": 2025, "publication_date": "2025-10-02", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5055842287", "display_name": "Yuhao Sun", "orcid": "https://orcid.org/0000-0001-6987-3721", "observed_orcids": ["https://orcid.org/0000-0001-6987-3721"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sun, Yuhao", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Xu, Zhuoer", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Zhuoer", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5039364160", "display_name": "Shiwen Cui", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cui, Shiwen", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5051206362", "display_name": "Kun Yang", "orcid": "https://orcid.org/0000-0001-7908-2675", "observed_orcids": ["https://orcid.org/0000-0001-7908-2675"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Kun", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5089799495", "display_name": "Lingyun Yu", "orcid": "https://orcid.org/0000-0001-6403-761X", "observed_orcids": ["https://orcid.org/0000-0001-6403-761X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Lingyun", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119846710", "display_name": "Yongdong Zhang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Yongdong", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5019450768", "display_name": "Hongtao Xie", "orcid": "https://orcid.org/0000-0001-6969-4897", "observed_orcids": ["https://orcid.org/0000-0001-6969-4897"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xie, Hongtao", "raw_affiliation_strings": ["University of Science and Technology of China , Independent researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Science and Technology of China , Independent researcher", "institution_ids": []}]}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs)": [3], "have": [4], "achieved": [5], "remarkable": [6], "progress": [7], "across": [8, 137], "a": [9, 56, 76, 86, 101, 122, 186], "wide": [10], "range": [11], "of": [12, 129], "tasks,": [13], "but": [14], "remain": [15], "vulnerable": [16], "to": [17, 105], "safety": [18, 29, 62, 96, 107, 123, 133, 150, 168], "risks": [19], "such": [20], "as": [21, 85], "harmful": [22, 153], "content": [23], "generation": [24], "and": [25, 37, 48, 72, 94, 134, 142, 154, 181, 199], "jailbreak": [26, 155], "attacks.": [27], "Existing": [28], "techniques": [30], "--": [31, 40], "including": [32], "external": [33], "guardrails,": [34], "inference-time": [35, 172], "guidance,": [36], "post-training": [38], "alignment": [39, 195], "each": [41], "face": [42], "limitations": [43], "in": [44], "balancing": [45], "safety,": [46], "utility,": [47], "controllability.": [49], "In": [50], "this": [51], "work,": [52], "we": [53, 120], "propose": [54], "UpSafe$^\\circ$C,": [55], "unified": [57], "framework": [58], "for": [59, 189], "enhancing": [60], "LLM": [61, 190], "through": [63], "safety-aware": [64], "upcycling.": [65], "Our": [66, 183], "approach": [67], "first": [68], "identifies": [69], "safety-critical": [70], "layers": [71], "upcycles": [73], "them": [74], "into": [75], "sparse": [77], "Mixture-of-Experts": [78], "(MoE)": [79], "structure,": [80], "where": [81], "the": [82, 130, 176], "router": [83], "acts": [84], "soft": [87], "guardrail": [88], "that": [89, 146, 167, 174], "selectively": [90], "activates": [91], "original": [92], "MLPs": [93], "added": [95], "experts.": [97], "We": [98], "further": [99], "introduce": [100, 121], "two-stage": [102], "SFT": [103], "strategy": [104], "strengthen": [106], "discrimination": [108], "while": [109, 157], "preserving": [110], "general": [111, 162], "capabilities.": [112], "To": [113], "enable": [114], "flexible": [115], "control": [116, 173], "at": [117], "inference": [118], "time,": [119], "temperature": [124, 169], "mechanism,": [125], "allowing": [126], "dynamic": [127], "adjustment": [128], "trade-off": [131], "between": [132, 179], "utility.": [135], "Experiments": [136], "multiple": [138], "benchmarks,": [139], "base": [140], "model,": [141], "model": [143], "scales": [144], "demonstrate": [145], "UpSafe$^\\circ$C": [147], "achieves": [148, 175], "robust": [149], "improvements": [151], "against": [152], "inputs,": [156], "maintaining": [158], "competitive": [159], "performance": [160], "on": [161], "tasks.": [163], "Moreover,": [164], "analysis": [165], "shows": [166], "provides": [170], "fine-grained": [171], "Pareto-optimal": [177], "frontier": [178], "utility": [180], "safety.": [182], "results": [184], "highlight": [185], "new": [187], "direction": [188], "safety:": [191], "moving": [192], "from": [193], "static": [194], "toward": [196], "dynamic,": [197], "modular,": [198], "inference-aware": [200], "control.": [201]}, "primary_topic": {"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.986299991607666, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.986299991607666, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.0038999998942017555, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10883", "display_name": "Ethics and Social Impacts of AI", "score": 0.0032999999821186066, "subfield": {"id": "https://openalex.org/subfields/3311", "display_name": "Safety Research"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.02194", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2510.02194", "pdf_url": "https://arxiv.org/pdf/2510.02194", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414836283", "doi": "https://doi.org/10.48550/arxiv.2502.15920", "title": "Self-Taught Agentic Long Context Understanding", "publication_year": 2025, "publication_date": "2025-02-21", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5052671723", "display_name": "Yufan Zhuang", "orcid": "https://orcid.org/0000-0003-4063-6460", "observed_orcids": ["https://orcid.org/0000-0003-4063-6460", "https://orcid.org/0009-0001-7506-7475"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhuang, Yufan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5139762370", "display_name": "Xiaodong Yu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Xiaodong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5115829690", "display_name": "Jialian Wu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Jialian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101181950", "display_name": "Ximeng Sun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sun, Ximeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5107953191", "display_name": "Ze Wang", "orcid": "https://orcid.org/0009-0000-4619-7943", "observed_orcids": ["https://orcid.org/0009-0000-4619-7943"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Ze", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100325209", "display_name": "Jiang Liu", "orcid": "https://orcid.org/0000-0001-6281-6505", "observed_orcids": ["https://orcid.org/0000-0001-6281-6505"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Jiang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5011372688", "display_name": "Yu-Sheng Su", "orcid": "https://orcid.org/0000-0002-1531-3363", "observed_orcids": ["https://orcid.org/0000-0002-1531-3363"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Su, Yusheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5039500313", "display_name": "Jingbo Shang", "orcid": "https://orcid.org/0000-0002-7249-4404", "observed_orcids": ["https://orcid.org/0000-0002-7249-4404"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Shang, Jingbo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101728120", "display_name": "Zicheng Liu", "orcid": "https://orcid.org/0000-0003-1106-2963", "observed_orcids": ["https://orcid.org/0000-0003-1106-2963"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Zicheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5027115167", "display_name": "Emad Barsoum", "orcid": "https://orcid.org/0000-0002-4097-8690", "observed_orcids": ["https://orcid.org/0000-0002-4097-8690"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Barsoum, Emad", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Answering": [0], "complex,": [1], "long-context": [2, 175, 187], "questions": [3, 66], "remains": [4], "a": [5, 28, 75, 82, 93, 101, 167], "major": [6], "challenge": [7], "for": [8, 122], "large": [9], "language": [10], "models": [11, 59, 155], "(LLMs)": [12], "as": [13, 74, 197], "it": [14], "requires": [15], "effective": [16, 140], "question": [17], "clarifications": [18, 158], "and": [19, 67, 100, 130, 143, 159, 164, 185], "context": [20, 162, 198], "retrieval.": [21], "We": [22], "propose": [23], "Agentic": [24], "Long-Context": [25], "Understanding": [26], "(AgenticLU),": [27], "framework": [29], "designed": [30], "to": [31, 98, 115, 138, 148, 156], "enhance": [32, 149], "an": [33, 47], "LLM's": [34], "understanding": [35, 62], "of": [36, 53, 96, 104, 111], "such": [37], "queries": [38], "by": [39, 126], "integrating": [40], "targeted": [41], "self-clarification": [42], "with": [43, 92], "contextual": [44, 69], "grounding": [45], "within": [46], "agentic": [48], "workflow.": [49], "At": [50], "the": [51, 108, 119, 127], "core": [52], "AgenticLU": [54, 154, 179], "is": [55], "Chain-of-Clarifications": [56], "(CoC),": [57], "where": [58, 78], "refine": [60], "their": [61], "through": [63], "self-generated": [64], "clarification": [65], "corresponding": [68], "groundings.": [70], "By": [71], "scaling": [72], "inference": [73, 169], "tree": [76], "search": [77, 94, 113], "each": [79, 123], "node": [80], "represents": [81], "CoC": [83, 128], "step,": [84], "we": [85, 117], "achieve": [86], "97.8%": [87], "answer": [88], "recall": [89], "on": [90], "NarrativeQA": [91], "depth": [95], "up": [97], "three": [99], "branching": [102], "factor": [103], "eight.": [105], "To": [106], "amortize": [107], "high": [109], "cost": [110], "this": [112], "process": [114], "training,": [116], "leverage": [118], "preference": [120, 146], "pairs": [121], "step": [124], "obtained": [125], "workflow": [129], "perform": [131], "two-stage": [132], "model": [133], "finetuning:": [134], "(1)": [135], "supervised": [136], "finetuning": [137], "learn": [139], "decomposition": [141], "strategies,": [142], "(2)": [144], "direct": [145], "optimization": [147], "reasoning": [150, 192], "quality.": [151], "This": [152], "enables": [153], "generate": [157], "retrieve": [160], "relevant": [161], "effectively": [163], "efficiently": [165], "in": [166], "single": [168], "pass.": [170], "Extensive": [171], "experiments": [172], "across": [173], "seven": [174], "tasks": [176], "demonstrate": [177], "that": [178], "significantly": [180], "outperforms": [181], "state-of-the-art": [182], "prompting": [183], "methods": [184], "specialized": [186], "LLMs,": [188], "achieving": [189], "robust": [190], "multi-hop": [191], "while": [193], "sustaining": [194], "consistent": [195], "performance": [196], "length": [199], "grows.": [200]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9404000043869019, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9404000043869019, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.025599999353289604, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.01600000075995922, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.15920", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.15920", "pdf_url": "https://arxiv.org/pdf/2502.15920", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4406945378", "doi": "https://doi.org/10.48550/arxiv.2501.16453", "title": "Detecting Zero-Day Attacks in Digital Substations via In-Context Learning", "publication_year": 2025, "publication_date": "2025-01-27", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5114513370", "display_name": "Faizan Manzoor", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I859038795", "display_name": "Virginia Tech", "ror": "https://ror.org/02smfhw86", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I859038795"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Manzoor, Faizan", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , Virginia Tech"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , Virginia Tech", "institution_ids": ["https://openalex.org/I859038795"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070497653", "display_name": "Vanshaj Khattar", "orcid": "https://orcid.org/0009-0006-0734-5804", "observed_orcids": ["https://orcid.org/0009-0006-0734-5804"]}, "institutions": [{"id": "https://openalex.org/I859038795", "display_name": "Virginia Tech", "ror": "https://ror.org/02smfhw86", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I859038795"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Khattar, Vanshaj", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , Virginia Tech"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , Virginia Tech", "institution_ids": ["https://openalex.org/I859038795"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015104981", "display_name": "A. D. M. A. Herath", "orcid": "https://orcid.org/0000-0002-9312-367X", "observed_orcids": ["https://orcid.org/0000-0002-9312-367X"]}, "institutions": [{"id": "https://openalex.org/I859038795", "display_name": "Virginia Tech", "ror": "https://ror.org/02smfhw86", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I859038795"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Herath, Akila", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , Virginia Tech"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , Virginia Tech", "institution_ids": ["https://openalex.org/I859038795"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Black, Clifton", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I206578319", "display_name": "Southern Company (United States)", "ror": "https://ror.org/02r0g0093", "country_code": "US", "type": "company", "lineage": ["https://openalex.org/I206578319"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Black, Clifton", "raw_affiliation_strings": ["Southern Company"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Southern Company", "institution_ids": ["https://openalex.org/I206578319"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Nielsen, Matthew C", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I1332737386", "display_name": "General Electric (United States)", "ror": "https://ror.org/013msgt25", "country_code": "US", "type": "company", "lineage": ["https://openalex.org/I1332737386"]}, {"id": "https://openalex.org/I4210134512", "display_name": "GE Global Research (United States)", "ror": "https://ror.org/03e06qt98", "country_code": "US", "type": "company", "lineage": ["https://openalex.org/I4210134512"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Nielsen, Matthew C", "raw_affiliation_strings": ["GE Global Research"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "GE Global Research", "institution_ids": ["https://openalex.org/I1332737386", "https://openalex.org/I4210134512"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5091327453", "display_name": "Junho Hong", "orcid": "https://orcid.org/0000-0001-5035-8260", "observed_orcids": ["https://orcid.org/0000-0001-5035-8260", "https://orcid.org/0000-0002-4904-7657"]}, "institutions": [{"id": "https://openalex.org/I4210130704", "display_name": "University of Michigan\u2013Dearborn", "ror": "https://ror.org/035wtm547", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I4210130704"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Hong, Junho", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , University of Michigan-Dearborn "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , University of Michigan-Dearborn ", "institution_ids": ["https://openalex.org/I4210130704"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5114240228", "display_name": "Chen-Ching Liu", "orcid": "https://orcid.org/0000-0002-8941-7958", "observed_orcids": ["https://orcid.org/0000-0002-8941-7958"]}, "institutions": [{"id": "https://openalex.org/I859038795", "display_name": "Virginia Tech", "ror": "https://ror.org/02smfhw86", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I859038795"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Liu, Chen-Ching", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , Virginia Tech"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , Virginia Tech", "institution_ids": ["https://openalex.org/I859038795"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101811842", "display_name": "Ming Jin", "orcid": "https://orcid.org/0000-0002-6833-4811", "observed_orcids": ["https://orcid.org/0000-0002-6833-4811"]}, "institutions": [{"id": "https://openalex.org/I859038795", "display_name": "Virginia Tech", "ror": "https://ror.org/02smfhw86", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I859038795"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Jin, Ming", "raw_affiliation_strings": ["Electrical and Computer Engineering Department , Virginia Tech"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Electrical and Computer Engineering Department , Virginia Tech", "institution_ids": ["https://openalex.org/I859038795"]}]}], "abstract_inverted_index": {"The": [0, 91], "occurrences": [1], "of": [2, 29, 79, 87, 107, 152], "cyber": [3], "attacks": [4, 32, 65, 100, 132], "on": [5, 115, 130], "the": [6, 26, 38, 74, 80, 83, 95, 116, 121, 134, 142, 153], "power": [7], "grids": [8], "have": [9, 50], "been": [10, 51], "increasing": [11], "every": [12, 19], "year,": [13], "with": [14], "novel": [15, 62], "attack": [16, 54, 109], "techniques": [17], "emerging": [18], "year.": [20], "In": [21], "this": [22], "paper,": [23], "we": [24], "address": [25], "critical": [27], "challenge": [28], "detecting": [30], "novel/zero-day": [31], "in": [33, 56], "digital": [34, 58, 150], "substations": [35, 151], "that": [36, 72, 108, 120], "employ": [37], "IEC-61850": [39, 57, 117], "communication": [40], "protocol.": [41], "While": [42], "many": [43], "heuristic": [44], "and": [45, 101, 148], "machine": [46], "learning": [47, 76], "(ML)-based": [48], "methods": [49], "proposed": [52, 122], "for": [53, 144], "detection": [55, 128], "substations,": [59], "generalization": [60], "to": [61, 97], "or": [63], "zero-day": [64, 99, 131], "remains": [66], "challenging.": [67], "We": [68], "propose": [69], "an": [70], "approach": [71, 93], "leverages": [73], "in-context": [75], "(ICL)": [77], "capability": [78], "transformer": [81], "architecture,": [82], "fundamental": [84], "building": [85, 145], "block": [86], "large": [88], "language": [89], "models.": [90], "ICL": [92], "enables": [94], "model": [96], "detect": [98], "learn": [102], "from": [103], "a": [104], "few": [105], "examples": [106], "without": [110], "explicit": [111], "retraining.": [112], "Our": [113], "experiments": [114], "dataset": [118], "demonstrate": [119], "method": [123], "achieves": [124], "more": [125, 146], "than": [126], "$85\\%$": [127], "accuracy": [129], "while": [133], "existing": [135], "state-of-the-art": [136], "baselines": [137], "fail.": [138], "This": [139], "work": [140], "paves": [141], "way": [143], "secure": [147], "resilient": [149], "future.": [154]}, "primary_topic": {"id": "https://openalex.org/T10917", "display_name": "Smart Grid Security and Resilience", "score": 0.9990000128746033, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10917", "display_name": "Smart Grid Security and Resilience", "score": 0.9990000128746033, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10400", "display_name": "Network Security and Intrusion Detection", "score": 0.0005000000237487257, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T14276", "display_name": "Power Systems and Technologies", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2501.16453", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2501.16453", "pdf_url": "https://arxiv.org/pdf/2501.16453", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417183093", "doi": "https://doi.org/10.48550/arxiv.2507.01652", "title": "Autoregressive Image Generation with Linear Complexity: A Spatial-Aware Decay Perspective", "publication_year": 2025, "publication_date": "2025-07-02", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5102015118", "display_name": "Yuxin Mao", "orcid": "https://orcid.org/0000-0002-2837-2290", "observed_orcids": ["https://orcid.org/0000-0002-2837-2290"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Mao, Yuxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Qin, Zhen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qin, Zhen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5047303316", "display_name": "Jinxing Zhou", "orcid": "https://orcid.org/0000-0001-6402-7593", "observed_orcids": ["https://orcid.org/0000-0001-6402-7593"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Jinxing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111614248", "display_name": "Hui Deng", "orcid": "https://orcid.org/0000-0002-8002-236X", "observed_orcids": ["https://orcid.org/0000-0002-8002-236X", "https://orcid.org/0000-0002-7116-7188", "https://orcid.org/0009-0003-5135-9590", "https://orcid.org/0009-0006-7611-7485"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Deng, Hui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063135412", "display_name": "Xuyang Shen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Shen, Xuyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101863814", "display_name": "Bin Fan", "orcid": "https://orcid.org/0000-0001-9191-8455", "observed_orcids": ["https://orcid.org/0000-0001-9191-8455"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Fan, Bin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100345515", "display_name": "Jing Zhang", "orcid": "https://orcid.org/0009-0002-4129-7237", "observed_orcids": ["https://orcid.org/0009-0002-4129-7237"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Jing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066621365", "display_name": "Yiran Zhong", "orcid": "https://orcid.org/0000-0003-1404-3610", "observed_orcids": ["https://orcid.org/0000-0003-1404-3610"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhong, Yiran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5036202579", "display_name": "Yuchao Dai", "orcid": "https://orcid.org/0000-0002-4432-7406", "observed_orcids": ["https://orcid.org/0000-0002-4432-7406"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dai, Yuchao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Autoregressive": [0], "(AR)": [1], "models": [2, 27], "have": [3, 61], "garnered": [4], "significant": [5], "attention": [6, 59, 102, 146], "in": [7, 66, 89], "image": [8, 77, 114, 141, 161], "generation": [9, 78, 162], "for": [10, 178], "their": [11, 82], "ability": [12], "to": [13, 50, 84, 147], "effectively": [14], "capture": [15, 85], "both": [16], "local": [17], "and": [18, 45, 164, 174], "global": [19], "structures": [20], "within": [21, 111], "visual": [22, 90], "data.": [23, 91], "However,": [24], "prevalent": [25], "AR": [26], "predominantly": [28], "rely": [29], "on": [30, 122, 133, 155], "the": [31, 51, 112, 168], "transformer": [32], "architectures,": [33], "which": [34], "are": [35], "beset": [36], "by": [37, 116], "quadratic": [38], "computational": [39, 165], "complexity": [40], "concerning": [41], "input": [42], "sequence": [43, 130], "length": [44], "substantial": [46], "memory": [47], "overhead": [48], "due": [49], "necessity": [52], "of": [53, 81], "maintaining": [54], "key-value": [55], "caches.": [56], "Although": [57], "linear": [58, 152, 171], "mechanisms": [60], "successfully": [62], "reduced": [63], "this": [64, 134], "burden": [65], "language": [67], "models,": [68], "our": [69], "initial": [70], "experiments": [71], "reveal": [72], "that": [73, 104, 143], "they": [74], "significantly": [75], "degrade": [76], "quality": [79], "because": [80], "inability": [83], "critical": [86], "long-range": [87], "dependencies": [88], "We": [92], "propose": [93], "Linear": [94], "Attention": [95], "with": [96, 151], "Spatial-Aware": [97], "Decay": [98], "(LASAD),": [99], "a": [100], "novel": [101], "mechanism": [103], "explicitly": [105], "preserves": [106], "genuine": [107], "2D": [108, 124], "spatial": [109, 125, 149, 175], "relationships": [110], "flattened": [113], "sequences": [115], "computing": [117], "position-dependent": [118], "decay": [119], "factors": [120], "based": [121], "true": [123], "location": [126], "rather": [127], "than": [128], "1D": [129], "positions.": [131], "Based": [132], "mechanism,": [135], "we": [136], "present": [137], "LASADGen,": [138], "an": [139], "autoregressive": [140], "generator": [142], "enables": [144], "selective": [145], "relevant": [148], "contexts": [150], "complexity.": [153], "Experiments": [154], "ImageNet": [156], "show": [157], "LASADGen": [158], "achieves": [159], "state-of-the-art": [160], "performance": [163], "efficiency,": [166], "bridging": [167], "gap": [169], "between": [170], "attention's": [172], "efficiency": [173], "understanding": [176], "needed": [177], "high-quality": [179], "generation.": [180]}, "primary_topic": {"id": "https://openalex.org/T10775", "display_name": "Generative Adversarial Networks and Image Synthesis", "score": 0.9961000084877014, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10775", "display_name": "Generative Adversarial Networks and Image Synthesis", "score": 0.9961000084877014, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0017999999690800905, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.0010999999940395355, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2507.01652", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2507.01652", "pdf_url": "https://arxiv.org/pdf/2507.01652", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417006854", "doi": "https://doi.org/10.48550/arxiv.2512.02556", "title": "DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models", "publication_year": 2025, "publication_date": "2025-12-02", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5093670579", "display_name": "DeepSeek-AI", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "DeepSeek-AI", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015078816", "display_name": "Aixin Liu", "orcid": "https://orcid.org/0000-0001-7152-511X", "observed_orcids": ["https://orcid.org/0000-0001-7152-511X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Aixin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Mei, Aoxue", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Mei, Aoxue", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Lin, Bangcai", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Bangcai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5018887016", "display_name": "Bing Xue", "orcid": "https://orcid.org/0000-0002-7396-9782", "observed_orcids": ["https://orcid.org/0000-0002-7396-9782"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xue, Bing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5005144178", "display_name": "Bingxuan Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Bingxuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109505733", "display_name": "Bingzheng Xu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Bingzheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103178710", "display_name": "Bowen Wu", "orcid": "https://orcid.org/0009-0005-8791-9689", "observed_orcids": ["https://orcid.org/0009-0005-8791-9689"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Bochao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5138967760", "display_name": "Bowei Zhang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Bowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109678287", "display_name": "Chaofan Lin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Chaofan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Dong, Chen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Chen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5090328003", "display_name": "Chengda Lu", "orcid": "https://orcid.org/0000-0002-9452-4053", "observed_orcids": ["https://orcid.org/0000-0002-9452-4053"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lu, Chengda", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5079470006", "display_name": "Chenggang Zhao", "orcid": "https://orcid.org/0009-0005-2297-0790", "observed_orcids": ["https://orcid.org/0009-0005-2297-0790"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhao, Chenggang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5084839131", "display_name": "Chengqi Deng", "orcid": "https://orcid.org/0009-0006-4777-2448", "observed_orcids": ["https://orcid.org/0009-0006-4777-2448"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Deng, Chengqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101012867", "display_name": "Chenhao Xu", "orcid": "https://orcid.org/0009-0004-8593-9857", "observed_orcids": ["https://orcid.org/0009-0004-8593-9857"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Chenhao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5019374810", "display_name": "Chong Ruan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ruan, Chong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020456783", "display_name": "Damai Dai", "orcid": "https://orcid.org/0009-0004-9714-7902", "observed_orcids": ["https://orcid.org/0009-0004-9714-7902"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dai, Damai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5060364305", "display_name": "Daya Guo", "orcid": "https://orcid.org/0009-0008-0822-1517", "observed_orcids": ["https://orcid.org/0009-0008-0822-1517"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Guo, Daya", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5044040772", "display_name": "Yang Dejian", "orcid": "https://orcid.org/0000-0002-6677-5365", "observed_orcids": ["https://orcid.org/0000-0002-6677-5365"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Dejian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101594616", "display_name": "Deli Chen", "orcid": "https://orcid.org/0000-0002-4593-9295", "observed_orcids": ["https://orcid.org/0000-0002-4593-9295"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Deli", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113091614", "display_name": "Erhang Li", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Erhang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5002273360", "display_name": "Fangqi Zhou", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Fangqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109672590", "display_name": "Fangyun Lin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Fangyun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Dai, Fucong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dai, Fucong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Hao, Guangbo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hao, Guangbo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5085396184", "display_name": "Guan-Ting Chen", "orcid": "https://orcid.org/0000-0002-0414-1412", "observed_orcids": ["https://orcid.org/0000-0002-0414-1412"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Guanting", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111738491", "display_name": "Guowei Li", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Guowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhang, H.", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, H.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101711729", "display_name": "Hanwei Xu", "orcid": "https://orcid.org/0000-0002-3970-0245", "observed_orcids": ["https://orcid.org/0000-0002-3970-0245"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Hanwei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5042878459", "display_name": "Hao Li", "orcid": "https://orcid.org/0000-0003-4693-8523", "observed_orcids": ["https://orcid.org/0000-0003-4693-8523"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Liang, Haofen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liang, Haofen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5114246441", "display_name": "Haoran Wei", "orcid": "https://orcid.org/0009-0007-6723-8282", "observed_orcids": ["https://orcid.org/0009-0007-6723-8282"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wei, Haoran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100615599", "display_name": "Haowei Zhang", "orcid": "https://orcid.org/0009-0000-9277-377X", "observed_orcids": ["https://orcid.org/0009-0000-9277-377X", "https://orcid.org/0000-0002-1985-1126"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Haowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049311422", "display_name": "Haowen Luo", "orcid": "https://orcid.org/0009-0007-6245-4330", "observed_orcids": ["https://orcid.org/0009-0007-6245-4330"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Luo, Haowen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020700222", "display_name": "Haozhe Ji", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ji, Haozhe", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066575939", "display_name": "Honghui Ding", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ding, Honghui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5072855790", "display_name": "Hongxuan Tang", "orcid": "https://orcid.org/0009-0000-3150-0671", "observed_orcids": ["https://orcid.org/0009-0000-3150-0671"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Hongxuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5016089490", "display_name": "Huanqi Cao", "orcid": "https://orcid.org/0000-0002-3870-106X", "observed_orcids": ["https://orcid.org/0000-0002-3870-106X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cao, Huanqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102602497", "display_name": "Huazuo Gao", "orcid": "https://orcid.org/0009-0009-8238-440X", "observed_orcids": ["https://orcid.org/0009-0009-8238-440X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gao, Huazuo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Qu, Hui", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qu, Hui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5077369376", "display_name": "Hui Zeng", "orcid": "https://orcid.org/0000-0002-4137-7424", "observed_orcids": ["https://orcid.org/0000-0002-4137-7424"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zeng, Hui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103278071", "display_name": "Jialiang Huang", "orcid": "https://orcid.org/0000-0002-8387-7320", "observed_orcids": ["https://orcid.org/0000-0002-8387-7320"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Jialiang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5104127081", "display_name": "Jiashi Li", "orcid": "https://orcid.org/0009-0007-0709-7692", "observed_orcids": ["https://orcid.org/0009-0007-0709-7692"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Jiashi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101299439", "display_name": "Jiaxin Xu", "orcid": "https://orcid.org/0009-0003-5321-3202", "observed_orcids": ["https://orcid.org/0009-0003-5321-3202"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Jiaxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Hu, Jiewen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hu, Jiewen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5003566311", "display_name": "Jingchang Chen", "orcid": "https://orcid.org/0000-0002-1378-898X", "observed_orcids": ["https://orcid.org/0000-0002-1378-898X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Jingchang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Xiang, Jingting", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xiang, Jingting", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113747152", "display_name": "Jingyang Yuan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yuan, Jingyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5007742979", "display_name": "Jingyuan Cheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cheng, Jingyuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102224739", "display_name": "Jinhua Zhu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Jinhua", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Ran, Jun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ran, Jun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101828204", "display_name": "Junguang Jiang", "orcid": "https://orcid.org/0000-0003-3099-3405", "observed_orcids": ["https://orcid.org/0000-0003-3099-3405"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jiang, Junguang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5026467215", "display_name": "Junjie Qiu", "orcid": "https://orcid.org/0009-0002-3606-9491", "observed_orcids": ["https://orcid.org/0009-0002-3606-9491"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qiu, Junjie", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100757073", "display_name": "Junlong Li", "orcid": "https://orcid.org/0009-0006-4192-413X", "observed_orcids": ["https://orcid.org/0009-0006-4192-413X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Junlong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063711526", "display_name": "Junxiao Song", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Song, Junxiao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102020714", "display_name": "Kai Dong", "orcid": "https://orcid.org/0000-0002-4652-6687", "observed_orcids": ["https://orcid.org/0000-0002-4652-6687"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Kai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Gao, Kaige", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gao, Kaige", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Guan, Kang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Guan, Kang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101379621", "display_name": "Kexin Huang", "orcid": "https://orcid.org/0009-0009-6557-3145", "observed_orcids": ["https://orcid.org/0009-0009-6557-3145"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Kexin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5000718123", "display_name": "Kexing Zhou", "orcid": "https://orcid.org/0000-0001-7635-3425", "observed_orcids": ["https://orcid.org/0000-0001-7635-3425"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Kexing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5023674199", "display_name": "Kezhao Huang", "orcid": "https://orcid.org/0009-0006-7273-0952", "observed_orcids": ["https://orcid.org/0009-0006-7273-0952"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Kezhao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Yu, Kuai", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Kuai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5034611488", "display_name": "Lean Wang", "orcid": "https://orcid.org/0009-0000-7676-8065", "observed_orcids": ["https://orcid.org/0009-0000-7676-8065"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Lean", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5068839714", "display_name": "Lecong Zhang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Lecong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100435806", "display_name": "Lei Wang", "orcid": "https://orcid.org/0000-0001-8639-4917", "observed_orcids": ["https://orcid.org/0000-0001-8639-4917"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Lei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103063573", "display_name": "Liang Zhao", "orcid": "https://orcid.org/0009-0002-3974-3680", "observed_orcids": ["https://orcid.org/0009-0002-3974-3680"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhao, Liang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5074614946", "display_name": "Liangsheng Yin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yin, Liangsheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101908065", "display_name": "Lihua Guo", "orcid": "https://orcid.org/0000-0002-4336-7190", "observed_orcids": ["https://orcid.org/0000-0002-4336-7190", "https://orcid.org/0000-0002-4916-3930"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Guo, Lihua", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5036387082", "display_name": "Lingxiao Luo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Luo, Lingxiao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113319915", "display_name": "L. R. Ma", "orcid": "https://orcid.org/0000-0001-9717-1508", "observed_orcids": ["https://orcid.org/0000-0001-9717-1508"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ma, Linwang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5030808218", "display_name": "Litong Wang", "orcid": "https://orcid.org/0009-0005-9586-0760", "observed_orcids": ["https://orcid.org/0009-0005-9586-0760"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Litong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101859634", "display_name": "Liyue Zhang", "orcid": "https://orcid.org/0009-0005-5871-182X", "observed_orcids": ["https://orcid.org/0009-0005-5871-182X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Liyue", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Di, M. S.", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Di, M. S.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101910351", "display_name": "Mingjun Xu", "orcid": "https://orcid.org/0009-0003-6834-0191", "observed_orcids": ["https://orcid.org/0009-0003-6834-0191"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, M. Y", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5087014857", "display_name": "Mingchuan Zhang", "orcid": "https://orcid.org/0000-0002-2523-1089", "observed_orcids": ["https://orcid.org/0000-0002-2523-1089"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Mingchuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015985030", "display_name": "Minghua Zhang", "orcid": "https://orcid.org/0000-0002-3607-7734", "observed_orcids": ["https://orcid.org/0000-0002-3607-7734"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Minghua", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5050171932", "display_name": "Minghui Tang", "orcid": "https://orcid.org/0009-0000-9629-4098", "observed_orcids": ["https://orcid.org/0009-0000-9629-4098"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Minghui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101619976", "display_name": "Mingxu Zhou", "orcid": "https://orcid.org/0009-0002-8479-2844", "observed_orcids": ["https://orcid.org/0009-0002-8479-2844"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Mingxu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5053679681", "display_name": "Panpan Huang", "orcid": "https://orcid.org/0009-0008-2875-0184", "observed_orcids": ["https://orcid.org/0009-0008-2875-0184"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Panpan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Cong, Peixin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cong, Peixin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5065659839", "display_name": "Peiyi Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Peiyi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5009091910", "display_name": "Qiancheng Wang", "orcid": "https://orcid.org/0000-0001-7896-983X", "observed_orcids": ["https://orcid.org/0000-0001-7896-983X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Qiancheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5018225836", "display_name": "Qihao Zhu", "orcid": "https://orcid.org/0000-0002-5566-4491", "observed_orcids": ["https://orcid.org/0000-0002-5566-4491"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Qihao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100700135", "display_name": "Qingyang Li", "orcid": "https://orcid.org/0000-0002-8178-2416", "observed_orcids": ["https://orcid.org/0000-0002-8178-2416", "https://orcid.org/0000-0002-0497-9573"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Qingyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5041745327", "display_name": "Qinyu Chen", "orcid": "https://orcid.org/0009-0005-9480-6164", "observed_orcids": ["https://orcid.org/0009-0005-9480-6164"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Qinyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100915390", "display_name": "Qiushi Du", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Du, Qiushi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5107918038", "display_name": "R. Xu", "orcid": "https://orcid.org/0000-0002-1928-1717", "observed_orcids": ["https://orcid.org/0000-0002-1928-1717"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Ruiling", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5029083198", "display_name": "Ruiqi Ge", "orcid": "https://orcid.org/0000-0002-8529-3070", "observed_orcids": ["https://orcid.org/0000-0002-8529-3070"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ge, Ruiqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5034818978", "display_name": "Ruisong Zhang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Ruisong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Pan, Ruizhe", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Pan, Ruizhe", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5024300524", "display_name": "Runji Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Runji", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Yin, Runqiu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yin, Runqiu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5078354080", "display_name": "Runxin Xu", "orcid": "https://orcid.org/0000-0003-1501-3579", "observed_orcids": ["https://orcid.org/0000-0003-1501-3579"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Runxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Shen, Ruomeng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Shen, Ruomeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhang, Ruoyu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Ruoyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Liu, S. H.", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, S. H.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101289776", "display_name": "Shanghao Lu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lu, Shanghao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5035849215", "display_name": "Shangyan Zhou", "orcid": "https://orcid.org/0009-0002-9696-1801", "observed_orcids": ["https://orcid.org/0009-0002-9696-1801"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Shangyan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5046105992", "display_name": "Shanhuang Chen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Shanhuang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5074233262", "display_name": "Shaofei Cai", "orcid": "https://orcid.org/0000-0002-1195-7276", "observed_orcids": ["https://orcid.org/0000-0002-1195-7276"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cai, Shaofei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0, 30], "introduce": [1, 31], "DeepSeek-V3.2,": [2], "a": [3, 56, 117], "model": [4, 44], "that": [5, 37, 121], "harmonizes": [6], "high": [7], "computational": [8, 40], "efficiency": [9], "with": [10, 83], "superior": [11], "reasoning": [12, 79, 111], "and": [13, 61, 77, 96, 139], "agent": [14], "performance.": [15], "The": [16], "key": [17], "technical": [18], "breakthroughs": [19], "of": [20], "DeepSeek-V3.2": [21, 65], "are": [22], "as": [23], "follows:": [24], "(1)": [25], "DeepSeek": [26], "Sparse": [27], "Attention": [28], "(DSA):": [29], "DSA,": [32], "an": [33], "efficient": [34], "attention": [35], "mechanism": [36], "substantially": [38], "reduces": [39], "complexity": [41], "while": [42], "preserving": [43], "performance": [45, 87], "in": [46, 88, 100, 137], "long-context": [47], "scenarios.": [48], "(2)": [49], "Scalable": [50], "Reinforcement": [51], "Learning": [52], "Framework:": [53], "By": [54], "implementing": [55], "robust": [57], "reinforcement": [58], "learning": [59], "protocol": [60], "scaling": [62], "post-training": [63], "compute,": [64], "performs": [66], "comparably": [67], "to": [68], "GPT-5.": [69], "Notably,": [70], "our": [71], "high-compute": [72], "variant,": [73], "DeepSeek-V3.2-Speciale,": [74], "surpasses": [75], "GPT-5": [76], "exhibits": [78], "proficiency": [80], "on": [81], "par": [82], "Gemini-3.0-Pro,": [84], "achieving": [85], "gold-medal": [86], "both": [89], "the": [90, 97], "2025": [91], "International": [92, 98], "Mathematical": [93], "Olympiad": [94, 99], "(IMO)": [95], "Informatics": [101], "(IOI).": [102], "(3)": [103], "Large-Scale": [104], "Agentic": [105], "Task": [106], "Synthesis": [107], "Pipeline:": [108], "To": [109], "integrate": [110], "into": [112], "tool-use": [113], "scenarios,": [114], "we": [115], "developed": [116], "novel": [118], "synthesis": [119], "pipeline": [120], "systematically": [122], "generates": [123], "training": [124], "data": [125], "at": [126], "scale.": [127], "This": [128], "methodology": [129], "facilitates": [130], "scalable": [131], "agentic": [132], "post-training,": [133], "yielding": [134], "substantial": [135], "improvements": [136], "generalization": [138], "instruction-following": [140], "robustness": [141], "within": [142], "complex,": [143], "interactive": [144], "environments.": [145]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8744999766349792, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8744999766349792, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.09860000014305115, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.011699999682605267, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.02556", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.02556", "pdf_url": "https://arxiv.org/pdf/2512.02556", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 8, "is_retracted": false}, {"id": "https://openalex.org/W4406522060", "doi": "https://doi.org/10.31219/osf.io/e7m8u", "title": "Semantic Distillation Through Recursive Neural Contextualisation in Large Language Models", "publication_year": 2025, "publication_date": "2025-01-17", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5115920339", "display_name": "Laurence Andrith", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Laurence Andrith", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5115920340", "display_name": "Benedict Ashford", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Benedict Ashford", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5115920341", "display_name": "Gabriella Kingsleigh", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gabriella Kingsleigh", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Edmund Nightingale", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Edmund Nightingale", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Achieving": [0], "coherent": [1], "and": [2, 44, 57, 82, 126, 185], "semantically": [3], "consistent": [4], "outputs": [5], "remains": [6], "a": [7, 29, 62, 108, 179], "significant": [8, 207], "challenge": [9], "in": [10, 78, 97, 111, 200, 203, 212], "extended": [11], "text": [12], "generation": [13], "tasks,": [14], "where": [15], "models": [16, 40], "often": [17], "struggle": [18], "with": [19, 85, 144, 157], "contextual": [20, 83], "retention": [21], "over": [22], "long": [23], "sequences.": [24], "Recursive": [25], "neural": [26, 176], "contextualisation": [27, 177], "offers": [28], "groundbreaking": [30], "approach": [31, 123], "that": [32, 104, 133], "refines": [33], "the": [34, 90, 119, 122, 134, 151, 168, 183, 194], "internal": [35], "representations": [36], "of": [37, 66, 121, 150, 171, 187, 197], "large": [38], "language": [39, 189, 214], "through": [41], "iterative": [42], "feedback": [43, 52], "dynamic": [45], "memory": [46], "enhancements.": [47], "The": [48, 147, 191], "framework": [49], "integrates": [50], "recursive": [51, 135, 175, 198], "pathways,": [53], "adaptive": [54], "gating": [55], "mechanisms,": [56, 174], "hierarchical": [58], "positional": [59], "embeddings,": [60], "enabling": [61], "more": [63], "effective": [64], "alignment": [65, 80], "semantic": [67, 79, 204], "relationships": [68], "across": [69, 124], "varied": [70], "linguistic": [71], "structures.": [72], "Quantitative": [73], "results": [74], "highlight": [75], "substantial": [76], "improvements": [77], "scores": [81], "coherence,": [84], "qualitative": [86], "analyses": [87], "further": [88], "illustrating": [89], "framework\u2019s": [91], "capacity": [92], "to": [93], "preserve": [94], "logical": [95], "consistency": [96], "complex,": [98], "multi-turn": [99], "interactions.": [100], "Layer-specific": [101], "evaluations": [102], "reveal": [103], "deeper": [105], "layers": [106], "play": [107], "pivotal": [109], "role": [110], "consolidating": [112], "long-term": [113], "dependencies,": [114], "while": [115], "multilingual": [116], "tests": [117], "confirm": [118], "adaptability": [120, 186], "languages": [125], "syntactic": [127], "variances.": [128], "Memory": [129], "efficiency": [130], "assessments": [131], "demonstrate": [132, 193], "architecture": [136], "scales": [137], "effectively,": [138], "maintaining": [139], "practicality": [140], "for": [141, 163, 181, 209], "real-world": [142], "applications": [143], "constrained": [145], "resources.": [146], "modular": [148], "nature": [149], "proposed": [152], "modifications": [153], "ensures": [154], "seamless": [155], "compatibility": [156], "existing": [158], "transformer": [159], "architectures,": [160], "fostering": [161], "opportunities": [162], "widespread": [164], "adoption.": [165], "By": [166], "addressing": [167], "core": [169], "limitations": [170], "traditional": [172], "attention": [173], "establishes": [178], "pathway": [180], "advancing": [182], "coherence": [184], "natural": [188, 213], "systems.": [190], "findings": [192], "transformative": [195], "potential": [196], "mechanisms": [199], "bridging": [201], "gaps": [202], "retention,": [205], "providing": [206], "implications": [208], "future": [210], "developments": [211], "processing.": [215]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9761000275611877, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9761000275611877, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.01759999990463257, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.0019000000320374966, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.31219/osf.io/e7m8u", "is_oa": true, "landing_page_url": "https://doi.org/10.31219/osf.io/e7m8u", "pdf_url": "https://osf.io/e7m8u/download", "source": null, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4409721863", "doi": "https://doi.org/10.2196/preprints.76433", "title": "Multi-Criteria Optimization of Language Models for HFpEF Symptom Detection in Spanish Electronic Health Records (Preprint)", "publication_year": 2025, "publication_date": "2025-04-23", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5002165918", "display_name": "Jacinto Mata V\u00e1zquez", "orcid": "https://orcid.org/0000-0001-5329-9622", "observed_orcids": ["https://orcid.org/0000-0001-5329-9622"]}, "institutions": [{"id": "https://openalex.org/I78880903", "display_name": "Universidad de Huelva", "ror": "https://ror.org/03a1kt624", "country_code": "ES", "type": "education", "lineage": ["https://openalex.org/I78880903"]}], "countries": ["ES"], "is_corresponding": false, "raw_author_name": "Jacinto Mata", "raw_affiliation_strings": ["I2C Research Group Universidad de Huelva Huelva ES", "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain"], "raw_orcid": "https://orcid.org/0000-0001-5329-9622", "affiliations": [{"raw_affiliation_string": "I2C Research Group Universidad de Huelva Huelva ES", "institution_ids": ["https://openalex.org/I78880903"]}, {"raw_affiliation_string": "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain", "institution_ids": ["https://openalex.org/I78880903"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5043282222", "display_name": "Victoria Pach\u00f3n", "orcid": "https://orcid.org/0000-0003-0697-4044", "observed_orcids": ["https://orcid.org/0000-0003-0697-4044"]}, "institutions": [{"id": "https://openalex.org/I78880903", "display_name": "Universidad de Huelva", "ror": "https://ror.org/03a1kt624", "country_code": "ES", "type": "education", "lineage": ["https://openalex.org/I78880903"]}], "countries": ["ES"], "is_corresponding": false, "raw_author_name": "Victoria Pach\u00f3n", "raw_affiliation_strings": ["I2C Research Group Universidad de Huelva Huelva ES", "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain"], "raw_orcid": "https://orcid.org/0000-0003-0697-4044", "affiliations": [{"raw_affiliation_string": "I2C Research Group Universidad de Huelva Huelva ES", "institution_ids": ["https://openalex.org/I78880903"]}, {"raw_affiliation_string": "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain", "institution_ids": ["https://openalex.org/I78880903"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5117273547", "display_name": "Ana Manovel", "orcid": "https://orcid.org/0000-0001-7015-170X", "observed_orcids": ["https://orcid.org/0000-0001-7015-170X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ana Manovel", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-7015-170X", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5079271501", "display_name": "Manuel Jes\u00fas Ma\u00f1a L\u00f3pez", "orcid": "https://orcid.org/0000-0002-7551-2401", "observed_orcids": ["https://orcid.org/0000-0002-7551-2401"]}, "institutions": [{"id": "https://openalex.org/I78880903", "display_name": "Universidad de Huelva", "ror": "https://ror.org/03a1kt624", "country_code": "ES", "type": "education", "lineage": ["https://openalex.org/I78880903"]}], "countries": ["ES"], "is_corresponding": false, "raw_author_name": "Manuel J. Ma\u00f1a", "raw_affiliation_strings": ["I2C Research Group Universidad de Huelva Huelva ES", "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain"], "raw_orcid": "https://orcid.org/0000-0002-7551-2401", "affiliations": [{"raw_affiliation_string": "I2C Research Group Universidad de Huelva Huelva ES", "institution_ids": ["https://openalex.org/I78880903"]}, {"raw_affiliation_string": "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain", "institution_ids": ["https://openalex.org/I78880903"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5102024368", "display_name": "Manuel de la Villa", "orcid": "https://orcid.org/0000-0003-3464-2944", "observed_orcids": ["https://orcid.org/0000-0003-3464-2944"]}, "institutions": [{"id": "https://openalex.org/I78880903", "display_name": "Universidad de Huelva", "ror": "https://ror.org/03a1kt624", "country_code": "ES", "type": "education", "lineage": ["https://openalex.org/I78880903"]}], "countries": ["ES"], "is_corresponding": false, "raw_author_name": "Manuel de la Villa", "raw_affiliation_strings": ["I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain"], "raw_orcid": "https://orcid.org/0000-0003-3464-2944", "affiliations": [{"raw_affiliation_string": "I\u00b2C Research Group , University of Huelva , Campus El Carmen s/n , Huelva , Spain", "institution_ids": ["https://openalex.org/I78880903"]}]}], "abstract_inverted_index": {"BACKGROUND": [0], "Heart": [1], "failure": [2, 106], "with": [3], "preserved": [4], "ejection": [5], "fraction": [6], "(HFpEF)": [7], "is": [8, 91], "a": [9, 17, 31, 56, 75, 191], "major": [10], "clinical": [11, 47, 98, 122, 195, 208, 244], "manifestation": [12], "of": [13, 63, 83, 96, 104, 109, 177, 181, 214, 226, 239], "cardiac": [14, 110], "amyloidosis": [15], "(CA),": [16], "condition": [18], "frequently": [19], "underdiagnosed": [20], "due": [21], "to": [22, 60, 92, 100, 148, 235], "its": [23], "nonspecific": [24], "symptomatology.": [25], "Electronic": [26], "health": [27], "records": [28], "(EHRs)": [29], "offer": [30], "promising": [32], "avenue": [33], "for": [34, 79, 194], "supporting": [35], "early": [36, 102], "symptom": [37], "detection": [38, 82], "through": [39], "natural": [40], "language": [41, 219], "processing": [42], "(NLP).": [43], "However,": [44], "identifying": [45], "relevant": [46, 150], "cues": [48], "within": [49], "unstructured": [50, 206], "narratives,": [51], "particularly": [52], "in": [53, 86, 211, 242], "Spanish,": [54], "remains": [55], "significant": [57], "challenge": [58], "owing": [59], "the": [61, 80, 94, 154, 212, 224, 237], "scarcity": [62], "annotated": [64, 125], "corpora": [65], "and": [66, 73, 126, 139, 158, 179, 217, 232], "domain-specific": [67, 228], "models.": [68], "OBJECTIVE": [69], "This": [70], "study": [71], "proposes": [72], "evaluates": [74], "Transformer-based": [76, 198], "NLP": [77, 240], "framework": [78], "automated": [81], "HFpEF-related": [84, 203], "symptoms": [85, 204], "Spanish": [87, 121, 207], "EHRs.": [88], "Our": [89, 221], "objective": [90], "assess": [93], "feasibility": [95], "leveraging": [97], "narratives": [99], "support": [101], "identification": [103], "heart": [105], "phenotypes": [107], "suggestive": [108], "amyloidosis.": [111], "METHODS": [112], "A": [113], "novel": [114], "corpus": [115], "was": [116], "developed": [117], "from": [118, 205], "over": [119], "15,000": [120], "documents,": [123], "manually": [124], "validated": [127], "by": [128], "cardiology": [129], "experts.": [130], "Several": [131], "Transformer": [132], "architectures": [133], "were": [134], "benchmarked,": [135], "including": [136], "general-purpose,": [137], "biomedical-specialized,": [138], "long-document": [140], "models": [141, 162, 184, 199], "(Longformers),": [142], "under": [143, 153], "three": [144], "optimization": [145, 234], "strategies": [146], "tailored": [147, 233], "clinically": [149], "metrics:": [151], "area": [152], "curve": [155], "(AUC),": [156], "F1-score,": [157], "sensitivity.": [159], "RESULTS": [160], "All": [161], "achieved": [163, 185], "strong": [164], "performance": [165, 238], "(AUC": [166], ">": [167], "0.940).": [168], "The": [169], "best-performing": [170], "model,": [171], "Longformer": [172], "Biomedical-Clinical,": [173], "reached": [174], "an": [175], "AUC": [176], "0.987": [178], "F1-score": [180], "0.985.": [182], "Sensitivity-optimized": [183], "false": [186], "negative": [187], "rates": [188], "below": [189], "3%,": [190], "critical": [192], "threshold": [193], "applicability.": [196], "CONCLUSIONS": [197], "can": [200], "robustly": [201], "detect": [202], "texts,": [209], "even": [210], "presence": [213], "class": [215], "imbalance": [216], "complex": [218], "patterns.": [220], "findings": [222], "highlight": [223], "importance": [225], "combining": [227], "pretraining,": [229], "long-context": [230], "modeling,": [231], "enhance": [236], "systems": [241], "high-impact": [243], "applications.": [245]}, "primary_topic": {"id": "https://openalex.org/T13702", "display_name": "Machine Learning in Healthcare", "score": 0.9442999958992004, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T13702", "display_name": "Machine Learning in Healthcare", "score": 0.9442999958992004, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11396", "display_name": "Artificial Intelligence in Healthcare", "score": 0.009800000116229057, "subfield": {"id": "https://openalex.org/subfields/3605", "display_name": "Health Information Management"}, "field": {"id": "https://openalex.org/fields/36", "display_name": "Health Professions"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.009399999864399433, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2196/preprints.76433", "is_oa": true, "landing_page_url": "https://doi.org/10.2196/preprints.76433", "pdf_url": "https://s3.ca-central-1.amazonaws.com/assets.jmir.org/assets/preprints/preprint-76433-submitted.pdf", "source": null, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416936453", "doi": "https://doi.org/10.48550/arxiv.2512.00418", "title": "Significant Other AI: Identity, Memory, and Emotional Regulation as Long-Term Relational Intelligence", "publication_year": 2025, "publication_date": "2025-11-29", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100400974", "display_name": "Sung\u2010Min Park", "orcid": "https://orcid.org/0000-0002-8359-8110", "observed_orcids": ["https://orcid.org/0000-0002-8359-8110"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Park, Sung", "raw_affiliation_strings": ["School of Data Science and Artificial Intelligence Taejae University , Seoul , Republic of Korea"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Data Science and Artificial Intelligence Taejae University , Seoul , Republic of Korea", "institution_ids": []}]}], "abstract_inverted_index": {"Significant": [0, 65], "Others": [1], "(SOs)": [2], "stabilize": [3], "identity,": [4], "regulate": [5], "emotion,": [6], "and": [7, 27, 49, 59, 80, 87, 101, 118, 136, 147], "support": [8, 38], "narrative": [9, 60, 99, 134], "meaning-making,": [10], "yet": [11], "many": [12, 162], "people": [13], "today": [14], "lack": [15, 164], "access": [16], "to": [17, 83], "such": [18], "relational": [19, 75, 115, 160], "anchors.": [20], "Recent": [21], "advances": [22], "in": [23], "large": [24], "language": [25], "models": [26], "memory-augmented": [28], "AI": [29, 155], "raise": [30], "the": [31, 159], "question": [32], "of": [33, 40, 74], "whether": [34, 154], "artificial": [35], "systems": [36], "could": [37], "some": [39], "these": [41], "functions.": [42], "Existing": [43], "empathic": [44], "AIs,": [45], "however,": [46], "remain": [47], "reactive": [48], "short-term,": [50], "lacking": [51], "autobiographical": [52], "memory,": [53, 96], "identity": [54, 93, 129], "modeling,": [55], "predictive": [56], "emotional": [57], "regulation,": [58], "coherence.": [61], "This": [62], "manuscript": [63], "introduces": [64], "Other": [66], "Artificial": [67], "Intelligence": [68], "(SO-AI)": [69], "as": [70, 143], "a": [71, 114, 119, 149], "new": [72], "domain": [73], "AI.": [76], "It": [77], "synthesizes": [78], "psychological": [79], "sociological": [81], "theory": [82], "define": [84], "SO": [85], "functions": [86], "derives": [88], "requirements": [89], "for": [90, 127, 152], "SO-AI,": [91], "including": [92], "awareness,": [94], "long-term": [95], "proactive": [97], "support,": [98], "co-construction,": [100], "ethical": [102], "boundary": [103], "enforcement.": [104], "A": [105, 122], "conceptual": [106], "architecture": [107], "is": [108], "proposed,": [109], "comprising": [110], "an": [111], "anthropomorphic": [112], "interface,": [113], "cognition": [116], "layer,": [117], "governance": [120], "layer.": [121], "research": [123], "agenda": [124], "outlines": [125], "methods": [126], "evaluating": [128], "stability,": [130], "longitudinal": [131], "interaction": [132], "patterns,": [133], "development,": [135], "sociocultural": [137], "impact.": [138], "SO-AI": [139], "reframes": [140], "AI-human": [141], "relationships": [142], "long-term,": [144], "identity-bearing": [145], "partnerships": [146], "provides": [148], "foundational": [150], "blueprint": [151], "investigating": [153], "can": [156], "responsibly": [157], "augment": [158], "stability": [161], "individuals": [163], "today.": [165]}, "primary_topic": {"id": "https://openalex.org/T10709", "display_name": "Social Robot Interaction and HRI", "score": 0.4253000020980835, "subfield": {"id": "https://openalex.org/subfields/3207", "display_name": "Social Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T10709", "display_name": "Social Robot Interaction and HRI", "score": 0.4253000020980835, "subfield": {"id": "https://openalex.org/subfields/3207", "display_name": "Social Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T10883", "display_name": "Ethics and Social Impacts of AI", "score": 0.3082999885082245, "subfield": {"id": "https://openalex.org/subfields/3311", "display_name": "Safety Research"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T14381", "display_name": "Psychiatry, Mental Health, Neuroscience", "score": 0.08760000020265579, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.00418", "is_oa": true, "landing_page_url": "https://arxiv.org/abs/2512.00418", "pdf_url": "https://arxiv.org/pdf/2512.00418", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4412409612", "doi": "https://doi.org/10.54254/2755-2721/2025.po24899", "title": "A Survey on Pre-trained Language Models Based on Deep Learning: Technological Development and Applications", "publication_year": 2025, "publication_date": "2025-07-10", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Yuansheng Lin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Yuansheng Lin", "raw_affiliation_strings": ["Capital University of Business, Beijing, China, 102218"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Capital University of Business, Beijing, China, 102218", "institution_ids": []}]}], "abstract_inverted_index": {"With": [0], "the": [1, 4, 9, 22, 41, 66, 83, 106, 116, 120, 127, 133], "advent": [2], "of": [3, 11, 24, 43, 69, 86, 97, 105, 109, 122, 129, 135, 145], "big": [5], "data": [6], "era": [7], "and": [8, 36, 54, 88, 94, 143, 156, 164], "enhancement": [10], "computing": [12], "capabilities,": [13], "deep": [14, 76], "learning": [15], "technologies": [16, 108], "have": [17, 38], "achieved": [18], "remarkable": [19], "breakthroughs": [20], "in": [21], "field": [23], "natural": [25], "language": [26, 31, 72, 130], "processing": [27], "(NLP).": [28], "Pre-trained": [29], "large": [30, 71], "models,": [32, 98, 124], "such": [33], "as": [34, 160, 162], "GPT": [35], "BERT,": [37], "significantly": [39, 125], "improved": [40], "performance": [42, 128], "various": [44], "NLP": [45], "tasks,": [46], "including": [47], "text": [48], "generation,": [49], "question-answering": [50], "systems,": [51], "sentiment": [52], "analysis,": [53], "machine": [55], "translation,": [56], "through": [57], "pre-training": [58, 84, 152], "on": [59, 75, 82], "large-scale": [60], "unsupervised": [61], "data.": [62], "This": [63], "paper": [64, 100], "reviews": [65], "latest": [67], "developments": [68], "pre-trained": [70, 110, 123], "models": [73], "based": [74], "learning,": [77], "with": [78], "a": [79, 91, 102], "particular": [80], "focus": [81], "methods": [85], "BERT": [87], "GPT.": [89], "Through": [90], "literature": [92], "review": [93], "comparative": [95], "analysis": [96], "this": [99], "provides": [101], "detailed": [103], "exploration": [104], "core": [107, 121], "models.": [111, 131], "The": [112], "study": [113], "finds": [114], "that": [115], "Transformer": [117], "architecture": [118], "is": [119], "enhancing": [126], "However,": [132], "expansion": [134], "model": [136, 154], "size": [137], "also": [138], "brings": [139], "increased": [140], "computational": [141], "costs": [142], "issues": [144], "interpretability.": [146], "Future": [147], "research": [148], "directions": [149], "include": [150], "efficient": [151], "methods,": [153], "compression": [155], "distillation,": [157], "multimodal": [158], "integration,": [159], "well": [161], "ethical": [163], "sustainability": [165], "issues.": [166]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9951000213623047, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9951000213623047, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.003599999938160181, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.00019999999494757503, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.54254/2755-2721/2025.po24899", "is_oa": true, "landing_page_url": "https://doi.org/10.54254/2755-2721/2025.po24899", "pdf_url": "https://www.ewadirect.com/proceedings/ace/article/view/24899/pdf", "source": {"id": "https://openalex.org/S4387281889", "display_name": "Applied and Computational Engineering", "issn_l": "2755-2721", "issn": ["2755-2721", "2755-273X"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core"], "host_organization": null, "host_organization_name": null, "host_organization_lineage": [], "host_organization_lineage_names": [], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Applied and Computational Engineering", "raw_type": "journal-article"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4415108702", "doi": "https://doi.org/10.48550/arxiv.2506.13356", "title": "StoryBench: A Dynamic Benchmark for Evaluating Long-Term Memory with Multi Turns", "publication_year": 2025, "publication_date": "2025-06-16", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Wan, Luanbo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wan, Luanbo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101524043", "display_name": "Weizhi Ma", "orcid": "https://orcid.org/0000-0001-5604-7527", "observed_orcids": ["https://orcid.org/0000-0001-5604-7527"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ma, Weizhi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Long-term": [0], "memory": [1, 38], "(LTM)": [2], "is": [3], "essential": [4], "for": [5], "large": [6], "language": [7], "models": [8, 135], "(LLMs)": [9], "to": [10, 33, 100, 120, 136, 158, 182], "achieve": [11], "autonomous": [12], "intelligence": [13], "in": [14, 21, 45, 54, 64, 188], "complex,": [15], "evolving": [16], "environments.": [17, 164], "Despite": [18], "increasing": [19], "efforts": [20], "memory-augmented": [22], "and": [23, 49, 53, 131, 140, 184], "retrieval-based": [24], "architectures,": [25], "there": [26], "remains": [27], "a": [28, 75, 154], "lack": [29], "of": [30, 59, 148, 170], "standardized": [31], "benchmarks": [32, 41], "systematically": [34], "evaluate": [35], "LLMs'": [36, 160], "long-term": [37], "abilities.": [39], "Existing": [40], "still": [42], "face": [43], "challenges": [44], "evaluating": [46], "knowledge": [47], "retention": [48], "dynamic": [50], "sequential": [51], "reasoning,": [52], "their": [55, 62], "own": [56], "flexibility,": [57], "all": [58], "which": [60], "limit": [61], "effectiveness": [63, 169], "assessing": [65], "models'": [66], "LTM": [67, 161, 187], "capabilities.": [68], "To": [69], "address": [70], "these": [71], "gaps,": [72], "we": [73, 151], "propose": [74], "novel": [76], "benchmark": [77, 115], "framework": [78], "based": [79], "on": [80], "interactive": [81], "fiction": [82], "games,": [83], "featuring": [84], "dynamically": [85], "branching": [86], "storylines": [87], "with": [88, 125], "complex": [89], "reasoning": [90, 122], "structures.": [91], "These": [92], "structures": [93], "simulate": [94], "real-world": [95], "scenarios": [96], "by": [97], "requiring": [98, 134], "LLMs": [99], "navigate": [101], "hierarchical": [102], "decision": [103], "trees,": [104], "where": [105], "each": [106], "choice": [107], "triggers": [108], "cascading": [109], "dependencies": [110], "across": [111], "multi-turn": [112], "interactions.": [113], "Our": [114], "emphasizes": [116], "two": [117], "distinct": [118], "settings": [119], "test": [121, 159], "complexity:": [123], "one": [124], "immediate": [126], "feedback": [127], "upon": [128], "incorrect": [129], "decisions,": [130], "the": [132, 168, 179], "other": [133], "independently": [137], "trace": [138], "back": [139], "revise": [141], "earlier": [142], "choices": [143], "after": [144], "failure.": [145], "As": [146], "part": [147], "this": [149], "benchmark,": [150], "also": [152], "construct": [153], "new": [155], "dataset": [156], "designed": [157], "within": [162], "narrative-driven": [163], "We": [165], "further": [166], "validate": [167], "our": [171], "approach": [172], "through": [173], "detailed": [174], "experiments.": [175], "Experimental": [176], "results": [177], "demonstrate": [178], "benchmark's": [180], "ability": [181], "robustly": [183], "reliably": [185], "assess": [186], "LLMs.": [189]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5059000253677368, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5059000253677368, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.1509999930858612, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10906", "display_name": "AI-based Problem Solving and Planning", "score": 0.0803999975323677, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.13356", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.13356", "pdf_url": "https://arxiv.org/pdf/2506.13356", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7116781830", "doi": "https://doi.org/10.3390/bdcc10010001", "title": "Transformer Encoder vs. Mamba SSM: Lightweight Architectures for Machining Stability-Induced Surface-Quality Categorization", "publication_year": 2025, "publication_date": "2025-12-19", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5024636209", "display_name": "Jeong Hoon Ko", "orcid": "https://orcid.org/0000-0002-4531-6805", "observed_orcids": ["https://orcid.org/0000-0002-4531-6805"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Jeong Hoon Ko", "raw_affiliation_strings": ["Precision Manufacturing, Taizhou Institute of Zhejiang University, 618, West Section of Shifu Avenue, Taizhou 318000, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Precision Manufacturing, Taizhou Institute of Zhejiang University, 618, West Section of Shifu Avenue, Taizhou 318000, China", "institution_ids": ["https://openalex.org/I76130692"]}]}], "abstract_inverted_index": {"Recent": [0], "advances": [1], "in": [2, 14, 31], "large": [3], "language": [4], "models": [5], "(LLMs)": [6], "have": [7], "revolutionized": [8], "many": [9], "domains;": [10], "however,": [11], "their": [12, 102, 130], "adoption": [13], "manufacturing": [15, 137], "remains": [16], "limited.": [17], "This": [18], "article": [19], "explores": [20], "the": [21, 70, 79, 86, 110, 117], "potential": [22], "of": [23, 63, 121], "state-of-the-art": [24], "AI": [25, 42, 124], "methodologies": [26], "for": [27, 44, 51, 135], "chatter-induced": [28], "surface-quality": [29], "categorization": [30], "machining.": [32], "While": [33], "LLMs": [34], "typically": [35], "demand": [36], "extremely": [37], "high": [38], "computational": [39], "resources": [40], "(FLOPs),": [41], "systems": [43], "chatter": [45], "recognition": [46], "must": [47], "instead": [48], "be": [49], "designed": [50], "efficient": [52], "classification": [53], "or": [54], "regression": [55], "on": [56], "structured": [57], "sensor": [58], "signals\u2014requiring": [59], "compact": [60], "architectures": [61], "capable": [62], "approximately\u20140.001": [64], "s": [65], "inference.": [66], "To": [67], "address": [68], "this,": [69], "present": [71], "study": [72], "applies": [73], "and": [74, 85, 109, 119, 132], "compares": [75], "two": [76], "representative": [77], "architectures:": [78], "transformer": [80], "encoder,": [81], "an": [82], "attention-based": [83], "model,": [84], "Mamba": [87], "SSM": [88], "(State-Space": [89], "Model),": [90], "a": [91], "non-attention-based": [92], "state-space": [93], "alternative.": [94], "Two": [95], "case": [96], "studies": [97], "are": [98, 126], "conducted": [99], "to": [100], "evaluate": [101], "performance:": [103], "one": [104], "using": [105], "machining-dynamics-based": [106], "simulation": [107], "data,": [108], "other": [111], "employing": [112], "actual": [113], "experimental": [114], "measurements.": [115], "Finally,": [116], "advantages": [118], "limitations": [120], "these": [122], "cutting-edge": [123], "approaches": [125], "critically": [127], "discussed,": [128], "emphasizing": [129], "suitability": [131], "practical": [133], "challenges": [134], "real-world": [136], "deployment.": [138]}, "primary_topic": {"id": "https://openalex.org/T10188", "display_name": "Advanced machining processes and optimization", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/2210", "display_name": "Mechanical Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10188", "display_name": "Advanced machining processes and optimization", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/2210", "display_name": "Mechanical Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T13113", "display_name": "Engineering Technology and Methodologies", "score": 0.0035000001080334187, "subfield": {"id": "https://openalex.org/subfields/2209", "display_name": "Industrial and Manufacturing Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10220", "display_name": "Machine Fault Diagnosis Techniques", "score": 0.002400000113993883, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.3390/bdcc10010001", "is_oa": true, "landing_page_url": "https://doi.org/10.3390/bdcc10010001", "pdf_url": "https://www.mdpi.com/2504-2289/10/1/1/pdf", "source": {"id": "https://openalex.org/S4210238752", "display_name": "Big Data and Cognitive Computing", "issn_l": "2504-2289", "issn": ["2504-2289"], "is_oa": true, "is_in_doaj": true, "is_core": true, "listed_in": ["cwts-core", "doaj", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310310987", "host_organization_name": "Multidisciplinary Digital Publishing Institute", "host_organization_lineage": ["https://openalex.org/P4310310987"], "host_organization_lineage_names": ["Multidisciplinary Digital Publishing Institute"], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Big Data and Cognitive Computing", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416373982", "doi": "https://doi.org/10.48550/arxiv.2510.04008", "title": "RACE Attention: A Strictly Linear-Time Attention Layer for Training on Outrageously Large Contexts", "publication_year": 2025, "publication_date": "2025-10-05", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5076233117", "display_name": "Sahil Joshi", "orcid": "https://orcid.org/0000-0003-3505-9363", "observed_orcids": ["https://orcid.org/0000-0003-3505-9363"]}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Joshi, Sahil", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020686363", "display_name": "Agniva Chowdhury", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Chowdhury, Agniva", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5120555334", "display_name": "Amar Kanakamedala", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Kanakamedala, Amar", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Singh, Ekam", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Singh, Ekam", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5078135146", "display_name": "E. Y. Tu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Tu, Evan", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5024993683", "display_name": "Anshumali Shrivastava", "orcid": "https://orcid.org/0000-0002-5042-2856", "observed_orcids": ["https://orcid.org/0000-0002-5042-2856"]}, "institutions": [{"id": "https://openalex.org/I74775410", "display_name": "Rice University", "ror": "https://ror.org/008zs3103", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I74775410"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Shrivastava, Anshumali", "raw_affiliation_strings": ["Department of Computer Science , Rice University , TX , USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , Rice University , TX , USA", "institution_ids": ["https://openalex.org/I74775410"]}]}], "abstract_inverted_index": {"Softmax": [0, 31, 68], "Attention": [1, 69, 81, 121, 194], "has": [2], "a": [3, 35, 40, 64, 87, 143, 148, 178, 197], "quadratic": [4], "time": [5, 135], "complexity": [6], "in": [7, 74, 177], "sequence": [8, 75], "length,": [9], "which": [10, 182], "becomes": [11], "prohibitive": [12], "to": [13, 67, 128, 157], "run": [14], "at": [15, 213], "long": [16], "contexts,": [17], "even": [18], "with": [19, 86], "highly": [20], "optimized": [21], "GPU": [22, 165], "kernels.": [23], "For": [24], "example,": [25], "FlashAttention-2/3": [26], "(exact,": [27], "GPU-optimized": [28], "implementations": [29], "of": [30, 39, 106, 155, 188], "Attention)": [32], "cannot": [33], "complete": [34], "single": [36, 41, 149, 179], "forward-backward": [37, 180], "pass": [38], "attention": [42, 93, 109, 150, 191], "layer": [43, 151], "once": [44], "the": [45, 83, 107, 186], "context": [46], "exceeds": [47], "~4": [48], "million": [49, 159, 168], "tokens": [50, 160, 169], "on": [51, 147, 161, 170, 206], "an": [52, 162, 171], "NVIDIA": [53, 163], "GH200": [54, 164], "(96": [55], "GB).": [56], "We": [57, 209], "introduce": [58], "Repeated": [59], "Arrays-of-Count": [60], "Estimators": [61], "(RACE)": [62], "Attention,": [63], "kernel-inspired": [65], "alternative": [66], "that": [70], "is": [71, 183], "strictly": [72], "linear": [73], "length": [76, 131], "and": [77, 91, 99, 117, 136, 152, 166, 199], "embedding": [78], "size.": [79], "RACE": [80, 120, 193], "replaces": [82], "exponential": [84], "kernel": [85], "sharpened": [88], "angular": [89], "similarity,": [90], "approximates": [92], "outputs": [94], "via": [95], "Gaussian": [96], "random": [97], "projections": [98], "soft": [100], "Locality-Sensitive": [101], "Hashing": [102], "(LSH),": [103], "avoiding": [104], "construction": [105], "full": [108], "matrix.": [110], "Across": [111], "language": [112, 115], "modeling,": [113, 116], "masked": [114], "text/image": [118], "classification,": [119], "matches": [122], "or": [123], "outperforms": [124], "strong": [125], "baselines": [126], "up": [127, 156], "64K": [129], "seqeuence": [130], "while": [132], "reducing": [133], "wall-clock": [134], "memory": [137], "usage.": [138], "In": [139], "addition,": [140], "we": [141], "conduct": [142], "controlled": [144], "scaling": [145], "study": [146], "demonstrate": [153], "processing": [154], "12": [158], "75": [167], "Intel": [172], "Xeon": [173], "Gold": [174], "5220R": [175], "CPU": [176], "pass,": [181], "well": [184], "beyond": [185], "capabilities": [187], "current": [189], "state-of-the-art": [190], "implementations.": [192], "thus": [195], "offers": [196], "practical": [198], "theoretically": [200], "grounded": [201], "mechanism": [202], "for": [203], "long-context": [204], "training": [205], "today's": [207], "hardware.": [208], "release": [210], "our": [211], "code": [212], "https://github.com/sahiljoshi515/RACE_Attention.": [214]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8730000257492065, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8730000257492065, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.044199999421834946, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.023399999365210533, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.04008", "is_oa": true, "landing_page_url": "https://arxiv.org/abs/2510.04008", "pdf_url": "https://arxiv.org/pdf/2510.04008", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416113188", "doi": "https://doi.org/10.48550/arxiv.2503.04359", "title": "LONGCODEU: Benchmarking Long-Context Language Models on Long Code Understanding", "publication_year": 2025, "publication_date": "2025-03-06", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100405683", "display_name": "Jia Li", "orcid": "https://orcid.org/0000-0002-3391-5720", "observed_orcids": ["https://orcid.org/0000-0002-3391-5720"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Jia", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Guo, Xuyuan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Guo, Xuyuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101722180", "display_name": "Lei Li", "orcid": "https://orcid.org/0000-0002-3517-4161", "observed_orcids": ["https://orcid.org/0000-0002-3517-4161", "https://orcid.org/0009-0005-6948-9435"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Lei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5069561378", "display_name": "Kechi Zhang", "orcid": "https://orcid.org/0000-0002-3290-0244", "observed_orcids": ["https://orcid.org/0000-0002-3290-0244"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Kechi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100447691", "display_name": "Ge Li", "orcid": "https://orcid.org/0000-0003-0140-0949", "observed_orcids": ["https://orcid.org/0000-0003-0140-0949", "https://orcid.org/0000-0003-4079-3968"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Ge", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100405683", "display_name": "Jia Li", "orcid": "https://orcid.org/0000-0002-3391-5720", "observed_orcids": ["https://orcid.org/0000-0002-3391-5720"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Jia", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5056146203", "display_name": "Zhengwei Tao", "orcid": "https://orcid.org/0000-0003-4025-6003", "observed_orcids": ["https://orcid.org/0000-0003-4025-6003"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tao, Zhengwei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5060027879", "display_name": "Fang Liu", "orcid": "https://orcid.org/0000-0002-3905-8133", "observed_orcids": ["https://orcid.org/0000-0002-3905-8133", "https://orcid.org/0009-0006-6517-3959"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Fang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5073065834", "display_name": "Chongyang Tao", "orcid": "https://orcid.org/0000-0002-4162-2119", "observed_orcids": ["https://orcid.org/0000-0002-4162-2119"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tao, Chongyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5027617553", "display_name": "Yuqi Sarah Zhu", "orcid": "https://orcid.org/0000-0003-1951-7053", "observed_orcids": ["https://orcid.org/0000-0003-1951-7053"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Yuqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5049100391", "display_name": "Zhi Gang Jin", "orcid": "https://orcid.org/0000-0003-1087-226X", "observed_orcids": ["https://orcid.org/0000-0003-1087-226X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jin, Zhi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Current": [0], "advanced": [1], "long-context": [2], "language": [3], "models": [4, 90], "offer": [5], "great": [6], "potential": [7], "for": [8, 32, 61, 105, 146, 153], "real-world": [9], "software": [10, 160], "engineering": [11], "applications.": [12], "However,": [13], "progress": [14], "in": [15, 101, 159], "this": [16, 38], "critical": [17], "domain": [18], "remains": [19], "hampered": [20], "by": [21], "a": [22, 28, 42], "fundamental": [23], "limitation:": [24], "the": [25, 110, 117, 135, 143], "absence": [26], "of": [27, 112, 128], "rigorous": [29], "evaluation": [30], "framework": [31], "long": [33, 43, 56, 76, 106, 118], "code": [34, 44, 57, 65, 77, 93, 107, 119], "understanding.": [35, 79, 108], "To": [36], "gap": [37], "obstacle,": [39], "we": [40], "propose": [41], "understanding": [45, 58, 141], "benchmark": [46], "LONGCODEU": [47, 86], "from": [48], "four": [49, 136], "aspects": [50], "(8": [51], "tasks)": [52], "to": [53], "evaluate": [54, 81], "LCLMs'": [55, 103], "ability": [59], "required": [60], "practical": [62], "applications,": [63], "including": [64], "unit": [66, 69, 72, 139], "perception,": [67], "intra-code": [68], "understanding,": [70, 74], "inter-code": [71, 138], "relation": [73, 140], "and": [75, 91, 156], "documentation": [78], "We": [80], "9": [82], "popular": [83], "LCLMs": [84, 113, 155], "on": [85], "(i.e.,": [87], "6": [88], "general": [89], "3": [92], "models).": [94], "Our": [95, 148], "experimental": [96], "results": [97], "reveal": [98], "key": [99], "limitations": [100], "current": [102], "capabilities": [104], "Particularly,": [109], "performance": [111], "drops": [114], "dramatically": [115], "when": [116], "length": [120], "is": [121, 142], "greater": [122], "than": [123], "32K,": [124], "falling": [125], "far": [126], "short": [127], "their": [129], "claimed": [130], "128K-1M": [131], "context": [132], "windows.": [133], "In": [134], "aspects,": [137], "most": [144], "challenging": [145], "LCLMs.": [147], "study": [149], "provides": [150], "valuable": [151], "insights": [152], "optimizing": [154], "driving": [157], "advancements": [158], "engineering.": [161]}, "primary_topic": {"id": "https://openalex.org/T10260", "display_name": "Software Engineering Research", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10260", "display_name": "Software Engineering Research", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 9.999999747378752e-05, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.04359", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.04359", "pdf_url": "https://arxiv.org/pdf/2503.04359", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4407759431", "doi": "https://doi.org/10.48550/arxiv.2502.12462", "title": "Emulating Retrieval Augmented Generation via Prompt Engineering for Enhanced Long Context Comprehension in LLMs", "publication_year": 2025, "publication_date": "2025-02-18", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Park, Joon", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Park, Joon", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5024780466", "display_name": "Kyohei Atarashi", "orcid": "https://orcid.org/0009-0007-7765-8002", "observed_orcids": ["https://orcid.org/0009-0007-7765-8002"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Atarashi, Kyohei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5088800219", "display_name": "Koh Takeuchi", "orcid": "https://orcid.org/0000-0002-3245-888X", "observed_orcids": ["https://orcid.org/0000-0002-3245-888X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Takeuchi, Koh", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5031707680", "display_name": "Hisashi Kashima", "orcid": "https://orcid.org/0000-0002-2770-0184", "observed_orcids": ["https://orcid.org/0000-0002-2770-0184"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kashima, Hisashi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"This": [0, 95], "paper": [1], "addresses": [2], "the": [3, 65, 69, 72, 164], "challenge": [4], "of": [5, 93, 129, 166], "comprehending": [6], "very": [7], "long": [8, 81], "contexts": [9], "in": [10, 40], "Large": [11], "Language": [12], "Models": [13], "(LLMs)": [14], "by": [15], "proposing": [16], "a": [17, 41, 80, 85, 195], "method": [18, 97], "that": [19, 179], "emulates": [20], "Retrieval": [21], "Augmented": [22], "Generation": [23], "(RAG)": [24], "through": [25], "specialized": [26], "prompt": [27, 161, 181], "engineering": [28], "and": [29, 71, 137, 154, 170, 192], "chain-of-thought": [30], "(CoT)": [31], "reasoning.": [32], "While": [33], "recent": [34], "LLMs": [35], "support": [36], "over": [37], "100,000": [38], "tokens": [39], "single": [42], "prompt,": [43], "simply": [44], "enlarging": [45], "context": [46], "windows": [47], "has": [48], "not": [49], "guaranteed": [50], "robust": [51], "multi-hop": [52], "reasoning": [53], "when": [54], "key": [55], "details": [56], "are": [57], "scattered": [58], "across": [59], "massive": [60], "input.": [61], "Our": [62], "approach": [63, 114, 142], "treats": [64], "model": [66], "as": [67, 149, 194], "both": [68], "retriever": [70], "reasoner:": [73], "it": [74], "first": [75], "tags": [76], "relevant": [77], "segments": [78], "within": [79], "passage,": [82], "then": [83], "employs": [84], "stepwise": [86], "CoT": [87], "workflow": [88], "to": [89, 133, 198], "integrate": [90], "these": [91], "pieces": [92], "evidence.": [94], "single-pass": [96], "thereby": [98], "reduces": [99], "reliance": [100], "on": [101, 108, 115], "an": [102], "external": [103], "retriever,": [104], "yet": [105], "maintains": [106], "focus": [107], "crucial": [109], "segments.": [110], "We": [111], "evaluate": [112], "our": [113, 141], "selected": [116], "tasks": [117], "from": [118], "BABILong,": [119], "which": [120], "interleaves": [121], "standard": [122], "bAbI": [123], "QA": [124], "problems": [125], "with": [126, 184], "large": [127], "amounts": [128], "distractor": [130], "text.": [131], "Compared": [132], "baseline": [134], "(no": [135], "retrieval)": [136], "naive": [138], "RAG": [139], "pipelines,": [140], "more": [143], "accurately": [144], "handles": [145], "multi-fact": [146], "questions": [147], "such": [148], "object": [150], "location": [151], "tracking,": [152], "counting,": [153], "indefinite": [155], "knowledge.": [156], "Furthermore,": [157], "we": [158], "analyze": [159], "how": [160], "structure,": [162], "including": [163], "order": [165], "question,": [167], "relevant-text": [168], "tags,": [169], "overall": [171], "instructions,": [172], "significantly": [173], "affects": [174], "performance.": [175], "These": [176], "findings": [177], "underscore": [178], "optimized": [180], "engineering,": [182], "combined": [183], "guided": [185], "reasoning,": [186], "can": [187], "enhance": [188], "LLMs'": [189], "long-context": [190], "comprehension": [191], "serve": [193], "lightweight": [196], "alternative": [197], "traditional": [199], "retrieval": [200], "pipelines.": [201]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9707000255584717, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9707000255584717, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.014800000004470348, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.006300000008195639, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.12462", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.12462", "pdf_url": "https://arxiv.org/pdf/2502.12462", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4408870195", "doi": "https://doi.org/10.3390/app15073608", "title": "Hybrid Transformer-Based Large Language Models for Word Sense Disambiguation in the Low-Resource Sesotho sa Leboa Language", "publication_year": 2025, "publication_date": "2025-03-25", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5078098302", "display_name": "Hlaudi Daniel Masethe", "orcid": "https://orcid.org/0000-0002-5538-3868", "observed_orcids": ["https://orcid.org/0000-0002-5538-3868"]}, "institutions": [{"id": "https://openalex.org/I137616099", "display_name": "Tshwane University of Technology", "ror": "https://ror.org/037mrss42", "country_code": "ZA", "type": "education", "lineage": ["https://openalex.org/I137616099"]}], "countries": ["ZA"], "is_corresponding": true, "raw_author_name": "Hlaudi Daniel Masethe", "raw_affiliation_strings": ["Department of Computer Science, Faculty of Information and Communication Technology, Tshwane University of Technology, Pretoria 0183, South Africa"], "raw_orcid": "https://orcid.org/0000-0002-5538-3868", "affiliations": [{"raw_affiliation_string": "Department of Computer Science, Faculty of Information and Communication Technology, Tshwane University of Technology, Pretoria 0183, South Africa", "institution_ids": ["https://openalex.org/I137616099"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5007875582", "display_name": "Mosima Anna Masethe", "orcid": "https://orcid.org/0000-0001-8825-7502", "observed_orcids": ["https://orcid.org/0000-0001-8825-7502"]}, "institutions": [{"id": "https://openalex.org/I2802393258", "display_name": "Sefako Makgatho Health Sciences University", "ror": "https://ror.org/003hsr719", "country_code": "ZA", "type": "education", "lineage": ["https://openalex.org/I2802393258"]}], "countries": ["ZA"], "is_corresponding": true, "raw_author_name": "Mosima Anna Masethe", "raw_affiliation_strings": ["Department of CSIT, School of Science and Technology, Sefako Makgatho Health Sciences University, Pretoria 0204, South Africa"], "raw_orcid": "https://orcid.org/0000-0001-8825-7502", "affiliations": [{"raw_affiliation_string": "Department of CSIT, School of Science and Technology, Sefako Makgatho Health Sciences University, Pretoria 0204, South Africa", "institution_ids": ["https://openalex.org/I2802393258"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5046107382", "display_name": "Sunday Olusegun Ojo", "orcid": "https://orcid.org/0000-0002-7197-8853", "observed_orcids": ["https://orcid.org/0000-0002-7197-8853"]}, "institutions": [{"id": "https://openalex.org/I177181097", "display_name": "Durban University of Technology", "ror": "https://ror.org/0303y7a51", "country_code": "ZA", "type": "education", "lineage": ["https://openalex.org/I177181097"]}], "countries": ["ZA"], "is_corresponding": false, "raw_author_name": "Sunday O. Ojo", "raw_affiliation_strings": ["Department of Information Technology, Faculty of Accounting and Informatics, Durban University of Technology, Durban 4001, South Africa"], "raw_orcid": "https://orcid.org/0000-0002-7197-8853", "affiliations": [{"raw_affiliation_string": "Department of Information Technology, Faculty of Accounting and Informatics, Durban University of Technology, Durban 4001, South Africa", "institution_ids": ["https://openalex.org/I177181097"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5013633670", "display_name": "Pius Adewale Owolawi", "orcid": "https://orcid.org/0000-0002-5781-5054", "observed_orcids": ["https://orcid.org/0000-0002-5781-5054"]}, "institutions": [{"id": "https://openalex.org/I137616099", "display_name": "Tshwane University of Technology", "ror": "https://ror.org/037mrss42", "country_code": "ZA", "type": "education", "lineage": ["https://openalex.org/I137616099"]}], "countries": ["ZA"], "is_corresponding": false, "raw_author_name": "Pius A. Owolawi", "raw_affiliation_strings": ["Department of Computer Engineering, Faculty of Information and Communication Technology, Tshwane University of Technology, Pretoria 0183, South Africa"], "raw_orcid": "https://orcid.org/0000-0002-5781-5054", "affiliations": [{"raw_affiliation_string": "Department of Computer Engineering, Faculty of Information and Communication Technology, Tshwane University of Technology, Pretoria 0183, South Africa", "institution_ids": ["https://openalex.org/I137616099"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5001227032", "display_name": "Fausto Giunchiglia", "orcid": "https://orcid.org/0000-0002-5903-6150", "observed_orcids": ["https://orcid.org/0000-0002-5903-6150"]}, "institutions": [{"id": "https://openalex.org/I193223587", "display_name": "University of Trento", "ror": "https://ror.org/05trd4x28", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I193223587"]}], "countries": ["IT"], "is_corresponding": false, "raw_author_name": "Fausto Giunchiglia", "raw_affiliation_strings": ["Department of Information Engineering and Computer Science, Faculty of Information Communication Technology, University of Trento, 38122 Trento, Italy"], "raw_orcid": "https://orcid.org/0000-0002-5903-6150", "affiliations": [{"raw_affiliation_string": "Department of Information Engineering and Computer Science, Faculty of Information Communication Technology, University of Trento, 38122 Trento, Italy", "institution_ids": ["https://openalex.org/I193223587"]}]}], "abstract_inverted_index": {"This": [0], "study": [1, 90], "addresses": [2], "a": [3, 52, 95, 102, 122], "lexical": [4, 33, 64], "ambiguity": [5, 34, 73], "issue": [6, 80], "in": [7, 35, 43, 83, 283], "Sesotho": [8, 36, 85], "sa": [9, 37, 86], "Leboa": [10, 38, 87], "that": [11, 324], "arises": [12], "from": [13], "terms": [14], "with": [15, 143, 181, 206, 215, 232, 275], "various": [16], "meanings,": [17], "often": [18], "known": [19], "as": [20, 157, 174], "homonyms": [21], "or": [22], "polysemous": [23], "words.": [24], "When": [25], "compared": [26], "to,": [27], "for": [28, 121, 311], "instance,": [29], "European": [30], "languages,": [31], "this": [32, 72, 89, 304], "causes": [39], "computational": [40], "semantic": [41], "problems": [42], "NLP": [44, 300], "when": [45], "trying": [46], "to": [47, 60, 71, 77, 93, 165, 201, 220, 295, 303], "identify": [48], "the": [49, 62, 79, 84, 112, 158, 167, 182, 187, 223, 246, 255, 268, 307], "lexicon": [50], "of": [51, 68, 81, 114, 145, 189, 211, 225, 249, 258, 278], "language.": [53], "In": [54, 75], "other": [55], "words,": [56], "it": [57], "is": [58], "challenging": [59], "determine": [61], "proper": [63], "category": [65], "and": [66, 107, 116, 131, 141, 150, 176, 197, 241, 251, 262, 292, 314, 319, 327, 341], "sense": [67, 97], "words": [69], "due": [70], "problem.": [74], "order": [76], "address": [78], "polysemy": [82], "language,": [88], "set": [91], "out": [92], "create": [94], "word": [96], "discrimination": [98], "(WSD)": [99], "scheme": [100], "using": [101], "corpus-based": [103], "hybrid": [104, 198, 343], "transformer-based": [105, 264], "architecture": [106, 164], "deep": [108], "learning": [109, 119], "models.": [110], "Additionally,": [111], "performance": [113, 298], "baseline": [115, 134, 169], "improved": [117, 219], "machine": [118], "models": [120, 135, 229], "sequence-based": [123], "natural": [124], "language": [125], "processing": [126], "(NLP)": [127], "task": [128], "was": [129], "assessed": [130], "compared.": [132], "The": [133, 191], "included": [136], "RNN-LSTM,": [137], "BiGRU,": [138, 204], "LSTMLM,": [139], "DeBERTa,": [140], "DistilBERT,": [142, 177], "accuracies": [144, 248], "61%,": [146], "79%,": [147], "74%,": [148], "70%,": [149, 279], "64%,": [151], "respectively.": [152], "Among": [153], "these,": [154], "BiGRU": [155, 214, 231, 238, 242, 270, 318], "emerged": [156], "strongest": [159], "performer,": [160], "leveraging": [161], "its": [162], "bidirectional": [163, 259], "achieve": [166], "highest": [168, 247], "accuracy.": [170, 190], "Transformer-based": [171], "models,": [172], "such": [173], "DeBERTa": [175, 240, 326], "demonstrated": [178, 235], "moderate": [179], "performance,": [180], "latter": [183], "prioritizing": [184], "efficiency": [185, 315], "at": [186], "cost": [188], "enhanced": [192], "results": [193], "explored": [194], "optimization": [195, 293], "techniques": [196], "model": [199, 273, 284, 344], "architectures": [200, 234], "improve": [202, 331], "performance.": [203], "optimized": [205], "ADAM,": [207], "achieved": [208, 245], "an": [209, 276], "accuracy": [210, 277, 313], "84%,": [212, 252], "while": [213], "attention": [216], "mechanisms": [217], "further": [218, 330], "85%,": [221], "showcasing": [222], "effectiveness": [224], "these": [226], "enhancements.": [227], "Hybrid": [228, 269], "integrating": [230], "transformer": [233], "varying": [236], "results.": [237], "+": [239, 243, 271], "ALBERT": [244], "85%": [250], "respectively,": [253], "highlighting": [254], "complementary": [256], "strengths": [257], "context": [260], "modeling": [261], "advanced": [263], "contextual": [265], "understanding.": [266], "Conversely,": [267], "RoBERTa": [272], "underperformed,": [274], "indicating": [280], "potential": [281], "mismatches": [282], "synergy.": [285], "These": [286], "findings": [287], "highlight": [288], "how": [289], "crucial": [290], "hybridization": [291], "are": [294, 316], "reaching": [296], "cutting-edge": [297], "on": [299, 337], "tasks.": [301], "According": [302], "study\u2019s": [305], "findings,": [306], "most": [308], "promising": [309], "approaches": [310], "fusing": [312], "attention-based": [317], "BiGRU\u2013transformer": [320], "hybrids,": [321], "especially": [322], "those": [323], "incorporate": [325], "ALBERT.": [328], "To": [329], "speed,": [332], "future": [333], "research": [334], "should": [335], "concentrate": [336], "exploring": [338], "task-specific": [339], "optimizations": [340], "improving": [342], "integration.": [345]}, "primary_topic": {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.9643999934196472, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.9643999934196472, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.03290000185370445, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10664", "display_name": "Sentiment Analysis and Opinion Mining", "score": 0.0006000000284984708, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.3390/app15073608", "is_oa": true, "landing_page_url": "https://doi.org/10.3390/app15073608", "pdf_url": null, "source": {"id": "https://openalex.org/S4210205812", "display_name": "Applied Sciences", "issn_l": "2076-3417", "issn": ["2076-3417"], "is_oa": true, "is_in_doaj": true, "is_core": true, "listed_in": ["cwts-core", "doaj", "norway-1"], "host_organization": "https://openalex.org/P4310310987", "host_organization_name": "Multidisciplinary Digital Publishing Institute", "host_organization_lineage": ["https://openalex.org/P4310310987"], "host_organization_lineage_names": ["Multidisciplinary Digital Publishing Institute"], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Applied Sciences", "raw_type": "journal-article"}, "cited_by_count": 6, "is_retracted": false}, {"id": "https://openalex.org/W4417433009", "doi": null, "title": "Beyond Isolated Capabilities: Bridging Long CoT Reasoning and Long-Context Understanding", "publication_year": 2025, "publication_date": "2025-07-20", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100365564", "display_name": "Yifei Wang", "orcid": "https://orcid.org/0000-0001-7382-5765", "observed_orcids": ["https://orcid.org/0000-0001-7382-5765", "https://orcid.org/0009-0001-1570-3871"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Wang, Yifei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Reasoning": [0], "distillation": [1, 23, 139], "has": [2, 170], "emerged": [3], "as": [4], "an": [5], "effective": [6], "approach": [7], "to": [8, 69], "enhance": [9], "the": [10, 18, 43, 67, 72, 162, 166], "reasoning": [11, 22, 97, 129, 150], "capabilities": [12], "of": [13, 20, 46, 56, 87], "smaller": [14], "language": [15], "models.": [16, 173], "However,": [17], "impact": [19], "large-scale": [21], "on": [24, 102], "other": [25], "critical": [26], "abilities,": [27], "particularly": [28, 40], "in-context": [29], "retrieval": [30], "and": [31, 54, 109, 119, 148, 155], "reasoning,": [32], "remains": [33], "unexplored.": [34], "This": [35, 158], "gap": [36], "in": [37, 107, 165], "understanding": [38], "is": [39], "significant": [41], "given": [42], "increasing": [44], "importance": [45], "Retrieval-Augmented": [47], "Generation": [48], "(RAG)": [49], "systems,": [50], "where": [51], "efficient": [52], "acquisition": [53], "utilization": [55], "contextual": [57], "information": [58, 112, 156], "are": [59], "paramount": [60], "for": [61, 94], "generating": [62], "reliable": [63], "responses.": [64], "Motivated": [65], "by": [66, 144], "need": [68], "understand": [70], "how": [71], "extended": [73, 114], "long-CoT": [74], "process": [75], "influences": [76], "long-context": [77, 133, 142, 172], "comprehension,": [78], "we": [79, 125], "conduct": [80], "a": [81, 85], "comprehensive": [82], "investigation": [83], "using": [84], "series": [86], "open-source": [88], "models": [89], "distilled": [90, 128], "from": [91, 113], "Deepseek-R1,": [92], "renowned": [93], "its": [95], "exceptional": [96], "capabilities.": [98], "Our": [99, 135], "study": [100], "focuses": [101], "evaluating": [103], "these": [104], "models'": [105], "performance": [106], "extracting": [108], "integrating": [110], "relevant": [111], "contexts": [115], "through": [116], "multi-document": [117], "question": [118], "answering": [120], "tasks.": [121], "Through": [122], "rigorous": [123], "experimentation,": [124], "demonstrate": [126], "that": [127, 138, 169], "patterns": [130], "significantly": [131], "improve": [132], "understanding.": [134], "analysis": [136, 154], "reveals": [137], "fosters": [140], "greater": [141], "awareness": [143], "promoting": [145], "more": [146], "detailed": [147], "explicit": [149], "processes": [151], "during": [152], "context": [153], "parsing.": [157], "advancement": [159], "effectively": [160], "mitigates": [161], "persistent": [163], "\"lost": [164], "middle\"": [167], "issue": [168], "hindered": [171]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9717000126838684, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9717000126838684, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.010400000028312206, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.007199999876320362, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2507.14849", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2507.14849", "pdf_url": "https://arxiv.org/pdf/2507.14849", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417028105", "doi": "https://doi.org/10.48550/arxiv.2512.03383", "title": "UniQL: Unified Quantization and Low-rank Compression for Adaptive Edge LLMs", "publication_year": 2025, "publication_date": "2025-12-03", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5082683969", "display_name": "Hung-Yueh Chiang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I86519309", "display_name": "The University of Texas at Austin", "ror": "https://ror.org/00hj54h04", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I86519309"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Chiang, Hung-Yueh", "raw_affiliation_strings": ["Chandra Family Department of Electrical and Computer Engineering , The University of Texas at Austin"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Chandra Family Department of Electrical and Computer Engineering , The University of Texas at Austin", "institution_ids": ["https://openalex.org/I86519309"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5005153980", "display_name": "Chi-Chih Chang", "orcid": "https://orcid.org/0000-0002-3154-1745", "observed_orcids": ["https://orcid.org/0000-0002-3154-1745"]}, "institutions": [{"id": "https://openalex.org/I205783295", "display_name": "Cornell University", "ror": "https://ror.org/05bnh6r87", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I205783295"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Chang, Chi-Chih", "raw_affiliation_strings": ["Department of Electrical and Computer Engineering , Cornell University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Electrical and Computer Engineering , Cornell University", "institution_ids": ["https://openalex.org/I205783295"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102900190", "display_name": "Yuchen Lu", "orcid": "https://orcid.org/0009-0004-5563-4905", "observed_orcids": ["https://orcid.org/0009-0004-5563-4905"]}, "institutions": [{"id": "https://openalex.org/I148366613", "display_name": "National Yang Ming Chiao Tung University", "ror": "https://ror.org/00se2k293", "country_code": "TW", "type": "education", "lineage": ["https://openalex.org/I148366613"]}], "countries": ["TW"], "is_corresponding": false, "raw_author_name": "Lu, Yu-Chen", "raw_affiliation_strings": ["Department of Computer Science , National Yang Ming Chiao Tung University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , National Yang Ming Chiao Tung University", "institution_ids": ["https://openalex.org/I148366613"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5037870921", "display_name": "Chien-Yu Lin", "orcid": "https://orcid.org/0000-0003-0118-3112", "observed_orcids": ["https://orcid.org/0000-0003-0118-3112"]}, "institutions": [{"id": "https://openalex.org/I4210100791", "display_name": "Paul G. Allen Family Foundation", "ror": "https://ror.org/01degd278", "country_code": "US", "type": "nonprofit", "lineage": ["https://openalex.org/I4210100791"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Lin, Chien-Yu", "raw_affiliation_strings": ["The Paul G. Allen School of Computer Science and Engineering , University of Washington"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Paul G. Allen School of Computer Science and Engineering , University of Washington", "institution_ids": ["https://openalex.org/I4210100791"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109999808", "display_name": "Kai\u2013Chiang Wu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I148366613", "display_name": "National Yang Ming Chiao Tung University", "ror": "https://ror.org/00se2k293", "country_code": "TW", "type": "education", "lineage": ["https://openalex.org/I148366613"]}], "countries": ["TW"], "is_corresponding": false, "raw_author_name": "Wu, Kai-Chiang", "raw_affiliation_strings": ["Department of Computer Science , National Yang Ming Chiao Tung University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science , National Yang Ming Chiao Tung University", "institution_ids": ["https://openalex.org/I148366613"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5104346656", "display_name": "Mohamed S. Abdelfattah", "orcid": "https://orcid.org/0000-0002-4568-8932", "observed_orcids": ["https://orcid.org/0000-0002-4568-8932"]}, "institutions": [{"id": "https://openalex.org/I205783295", "display_name": "Cornell University", "ror": "https://ror.org/05bnh6r87", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I205783295"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Abdelfattah, Mohamed S.", "raw_affiliation_strings": ["Department of Electrical and Computer Engineering , Cornell University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Electrical and Computer Engineering , Cornell University", "institution_ids": ["https://openalex.org/I205783295"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5065985595", "display_name": "Diana Marculescu", "orcid": "https://orcid.org/0000-0002-5734-4221", "observed_orcids": ["https://orcid.org/0000-0002-5734-4221"]}, "institutions": [{"id": "https://openalex.org/I86519309", "display_name": "The University of Texas at Austin", "ror": "https://ror.org/00hj54h04", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I86519309"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Marculescu, Diana", "raw_affiliation_strings": ["Chandra Family Department of Electrical and Computer Engineering , The University of Texas at Austin"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Chandra Family Department of Electrical and Computer Engineering , The University of Texas at Austin", "institution_ids": ["https://openalex.org/I86519309"]}]}], "abstract_inverted_index": {"Deploying": [0], "large": [1], "language": [2], "models": [3, 84, 163, 183, 196, 204], "(LLMs)": [4], "on": [5], "mobile": [6], "platforms": [7], "faces": [8], "significant": [9], "challenges": [10], "due": [11], "to": [12, 40, 85, 113, 154], "the": [13, 21, 35, 41, 141, 181], "limited": [14], "memory": [15, 166], "and": [16, 53, 73, 82, 122, 138, 161, 170, 190, 194, 198, 202], "shared": [17], "computational": [18], "resources": [19], "of": [20, 43, 168, 174, 180], "device.": [22], "Resource": [23], "availability": [24], "may": [25], "be": [26], "an": [27, 97], "issue": [28], "as": [29], "it": [30], "is": [31, 66], "directly": [32], "impacted": [33], "by": [34, 106], "current": [36], "device": [37], "workload,": [38], "adding": [39], "uncertainty": [42], "model": [44], "deployment.": [45], "We": [46], "introduce": [47, 96], "UniQL,": [48], "a": [49, 67, 123, 144, 165, 171], "unified": [50], "post-training": [51], "quantization": [52, 72, 115, 139], "low-rank": [54, 74], "compression": [55, 75], "framework": [56, 69, 134], "with": [57], "on-device": [58, 149], "configurable": [59, 150], "pruning": [60, 151, 186], "rates": [61, 152], "for": [62, 76, 120, 130], "edge": [63, 88], "LLMs.": [64], "UniQL": [65], "general": [68], "that": [70, 102, 159], "integrates": [71], "Transformers,": [77], "State": [78], "Space": [79], "Models": [80], "(SSMs),": [81], "hybrid": [83, 195], "support": [86], "diverse": [87], "applications.": [89], "In": [90], "our": [91], "proposed": [92], "joint": [93], "framework,": [94], "we": [95], "efficient": [98], "structured": [99], "weight-sorting": [100], "method": [101], "speeds": [103], "up": [104, 153], "computation": [105], "20x,": [107], "quantization-aware": [108], "singular": [109], "value": [110], "decomposition": [111], "(SVD)": [112], "minimize": [114], "errors,": [116], "state-aware": [117], "weight": [118], "sorting": [119], "SSMs,": [121], "fused": [124], "rotary": [125], "positional": [126], "embedding": [127], "(RoPE)": [128], "kernel": [129], "pruned": [131, 162], "models.": [132], "Our": [133, 156], "performs": [135], "weight-sorting,": [136], "fine-tuning,": [137], "in": [140, 143], "cloud": [142], "single-pass": [145], "workflow,": [146], "while": [147], "enabling": [148], "35%.": [155], "experiments": [157], "show": [158], "quantized": [160, 203], "achieve": [164], "reduction": [167], "4x-5.7x": [169], "token-throughput": [172], "improvement": [173], "2.7x-3.4x,": [175], "maintaining": [176], "accuracy": [177], "within": [178], "5%": [179], "original": [182], "at": [184], "15%": [185], "across": [187], "Transformers": [188], "(Llama3": [189], "Qwen2.5),": [191], "SSMs": [192], "(Mamba2),": [193], "(Nemotron-H": [197], "Bamba-v2).": [199], "The": [200], "code": [201], "are": [205], "available": [206], "at:": [207], "https://github.com/enyac-group/UniQL.": [208]}, "primary_topic": {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.0024999999441206455, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10273", "display_name": "IoT and Edge/Fog Computing", "score": 0.0024999999441206455, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.03383", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.03383", "pdf_url": "https://arxiv.org/pdf/2512.03383", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415102906", "doi": "https://doi.org/10.48550/arxiv.2503.09445", "title": "Astrea: A MOE-based Visual Understanding Model with Progressive Alignment", "publication_year": 2025, "publication_date": "2025-03-12", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5040461965", "display_name": "Xiaoda Yang", "orcid": "https://orcid.org/0009-0002-7297-4536", "observed_orcids": ["https://orcid.org/0009-0002-7297-4536"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Xiaoda", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5028459804", "display_name": "JunYu Lu", "orcid": "https://orcid.org/0009-0001-6239-0660", "observed_orcids": ["https://orcid.org/0009-0001-6239-0660"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lu, JunYu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101218526", "display_name": "Hongshun Qiu", "orcid": "https://orcid.org/0009-0009-3344-3813", "observed_orcids": ["https://orcid.org/0009-0009-3344-3813"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qiu, Hongshun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Li, Sijing", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Sijing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100348623", "display_name": "Hao Li", "orcid": "https://orcid.org/0000-0002-6652-4793", "observed_orcids": ["https://orcid.org/0000-0002-6652-4793"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Ji, Shengpeng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ji, Shengpeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Tang, Xudong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Xudong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Xu, Jiayang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Jiayang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Duan, Jiaqi", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Duan, Jiaqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5073215052", "display_name": "Ziyue Jiang", "orcid": "https://orcid.org/0009-0005-1358-8098", "observed_orcids": ["https://orcid.org/0009-0005-1358-8098"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jiang, Ziyue", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5035508811", "display_name": "Cong Lin", "orcid": "https://orcid.org/0000-0002-1567-1398", "observed_orcids": ["https://orcid.org/0000-0002-1567-1398"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Cong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5013362450", "display_name": "Sihang Cai", "orcid": "https://orcid.org/0009-0007-8693-7142", "observed_orcids": ["https://orcid.org/0009-0007-8693-7142"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cai, Sihang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5114126142", "display_name": "Zejian Xie", "orcid": "https://orcid.org/0009-0003-2967-0526", "observed_orcids": ["https://orcid.org/0009-0003-2967-0526"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xie, Zejian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101298759", "display_name": "Zhuoyang Song", "orcid": "https://orcid.org/0009-0001-7013-3488", "observed_orcids": ["https://orcid.org/0009-0001-7013-3488"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Song, Zhuoyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5080429423", "display_name": "Songxin Zhang", "orcid": "https://orcid.org/0009-0004-0110-6189", "observed_orcids": ["https://orcid.org/0009-0004-0110-6189"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Songxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Vision-Language": [0], "Models": [1], "(VLMs)": [2], "based": [3, 72], "on": [4, 73], "Mixture-of-Experts": [5], "(MoE)": [6], "architectures": [7], "have": [8], "emerged": [9], "as": [10], "a": [11, 18, 66, 97], "pivotal": [12], "paradigm": [13], "in": [14, 38], "multimodal": [15, 210], "understanding,": [16], "offering": [17], "powerful": [19], "framework": [20, 141], "for": [21, 146, 154, 207], "integrating": [22], "visual": [23, 44, 103], "and": [24, 31, 59, 150, 169], "linguistic": [25], "information.": [26], "However,": [27], "the": [28, 119, 188], "increasing": [29], "complexity": [30], "diversity": [32], "of": [33, 183], "tasks": [34, 164], "present": [35], "significant": [36], "challenges": [37], "coordinating": [39], "load": [40, 61], "balancing": [41], "across": [42, 161], "heterogeneous": [43, 83], "experts,": [45], "where": [46], "optimizing": [47], "one": [48], "specialist's": [49], "performance": [50, 181], "often": [51], "compromises": [52], "others'": [53], "capabilities.": [54], "To": [55], "address": [56], "task": [57, 200], "heterogeneity": [58, 201], "expert": [60, 84, 99, 156], "imbalance,": [62], "we": [63], "propose": [64], "Astrea,": [65], "novel": [67], "multi-expert": [68], "collaborative": [69], "VLM": [70, 120], "architecture": [71], "progressive": [74, 113, 193], "pre-alignment.": [75], "Astrea": [76], "introduces": [77], "three": [78], "key": [79], "innovations:": [80], "1)": [81], "A": [82, 107], "coordination": [85], "mechanism": [86], "that": [87, 192], "integrates": [88], "four": [89], "specialized": [90], "models": [91], "(detection,": [92], "segmentation,": [93], "classification,": [94], "captioning)": [95], "into": [96], "comprehensive": [98], "matrix": [100], "covering": [101], "essential": [102], "comprehension": [104], "elements;": [105], "2)": [106], "dynamic": [108], "knowledge": [109, 135], "fusion": [110], "strategy": [111], "featuring": [112], "pre-alignment": [114, 194], "to": [115, 133, 198], "harmonize": [116], "experts": [117], "within": [118], "latent": [121], "space": [122], "through": [123], "contrastive": [124, 144], "learning,": [125], "complemented": [126], "by": [127], "probabilistically": [128], "activated": [129], "stochastic": [130], "residual": [131], "connections": [132], "preserve": [134], "continuity;": [136], "3)": [137], "An": [138], "enhanced": [139], "optimization": [140], "utilizing": [142], "momentum": [143], "learning": [145], "long-range": [147], "dependency": [148], "modeling": [149], "adaptive": [151], "weight": [152], "allocators": [153], "real-time": [155], "contribution": [157], "calibration.": [158], "Extensive": [159], "evaluations": [160], "12": [162], "benchmark": [163], "spanning": [165], "VQA,": [166], "image": [167], "captioning,": [168], "cross-modal": [170], "retrieval": [171], "demonstrate": [172], "Astrea's": [173], "superiority": [174], "over": [175], "state-of-the-art": [176], "models,": [177], "achieving": [178], "an": [179], "average": [180], "gain": [182], "+4.7\\%.": [184], "This": [185], "study": [186], "provides": [187], "first": [189], "empirical": [190], "demonstration": [191], "strategies": [195], "enable": [196], "VLMs": [197], "overcome": [199], "limitations,": [202], "establishing": [203], "new": [204], "methodological": [205], "foundations": [206], "developing": [208], "general-purpose": [209], "agents.": [211]}, "primary_topic": {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.09445", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.09445", "pdf_url": "https://arxiv.org/pdf/2503.09445", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416339433", "doi": "https://doi.org/10.48550/arxiv.2511.10676", "title": "Pre-Attention Expert Prediction and Prefetching for Mixture-of-Experts Large Language Models", "publication_year": 2025, "publication_date": "2025-11-10", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Zhu, Shien", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Shien", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Bohl, Samuel", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bohl, Samuel", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5120361850", "display_name": "Robin Oester", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Oester, Robin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Alonso, Gustavo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Alonso, Gustavo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Mixture-of-Experts": [0], "(MoE)": [1], "Large": [2], "Language": [3], "Models": [4], "(LLMs)": [5], "efficiently": [6], "scale-up": [7], "the": [8, 23, 36, 45, 48, 58, 106, 120, 123, 127, 146, 176], "model": [9], "while": [10], "keeping": [11], "relatively": [12], "low": [13, 54], "inference": [14], "cost.": [15], "As": [16], "MoE": [17], "models": [18], "only": [19], "activate": [20], "part": [21], "of": [22, 108], "experts,": [24], "related": [25], "work": [26], "has": [27], "proposed": [28], "expert": [29, 83, 90, 152], "prediction": [30, 72, 84], "and": [31, 56, 88, 134, 164], "caching": [32], "methods": [33], "to": [34, 85, 137], "prefetch": [35], "experts": [37, 110], "for": [38, 51, 70], "faster": [39], "inference.": [40], "However,": [41], "existing": [42], "approaches": [43], "utilize": [44, 119], "activations": [46, 121], "from": [47], "previous": [49], "layer": [50, 60, 129], "prediction,": [52, 140], "incurring": [53], "accuracy": [55, 156, 174], "leave": [57], "first": [59, 147], "unoptimized.": [61], "Applying": [62], "complex": [63], "layers": [64], "or": [65], "even": [66], "training": [67], "standalone": [68], "networks": [69], "better": [71], "introduces": [73], "high": [74], "computation": [75], "overhead.": [76], "In": [77], "this": [78], "paper,": [79], "we": [80, 118], "propose": [81], "pre-attention": [82, 151], "achieve": [86, 138, 154], "accurate": [87, 139], "lightweight": [89], "prefetching.": [91], "The": [92], "key": [93], "insight": [94], "is": [95, 115], "that": [96, 104], "some": [97], "functions": [98, 114, 133], "in": [99, 126, 145], "LLMs": [100], "are": [101], "ranking-preserving,": [102], "indicating": [103], "matching": [105], "ranking": [107], "selected": [109], "using": [111], "simple": [112], "linear": [113, 132], "possible.": [116], "Therefore,": [117], "before": [122], "attention": [124], "block": [125], "same": [128], "with": [130], "2": [131], "ranking-aware": [135], "loss": [136], "which": [141], "also": [142], "supports": [143], "prefetching": [144], "layer.": [148], "Our": [149], "lightweight,": [150], "routers": [153], "93.03%": [155], "on": [157, 162, 166, 172], "DeepSeek": [158], "V2": [159], "Lite,": [160], "94.69%": [161], "Qwen3-30B,": [163], "97.62%": [165], "Phi-mini-MoE,": [167], "showing": [168], "about": [169], "15%": [170], "improvement": [171], "absolute": [173], "over": [175], "state-of-the-art": [177], "methods.": [178]}, "primary_topic": {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.7271000146865845, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.7271000146865845, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.1054999977350235, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.05900000035762787, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2511.10676", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2511.10676", "pdf_url": "https://arxiv.org/pdf/2511.10676", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414762093", "doi": "https://doi.org/10.48550/arxiv.2509.05165", "title": "KVCompose: Efficient Structured KV Cache Compression with Composite Tokens", "publication_year": 2025, "publication_date": "2025-09-05", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5119820576", "display_name": "Dmitry Akulov", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Akulov, Dmitry", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5028270135", "display_name": "Mohamed Sana", "orcid": "https://orcid.org/0000-0001-5644-5840", "observed_orcids": ["https://orcid.org/0000-0001-5644-5840"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sana, Mohamed", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045740906", "display_name": "Antonio De Domenico", "orcid": "https://orcid.org/0000-0003-1229-4045", "observed_orcids": ["https://orcid.org/0000-0003-1229-4045"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "De Domenico, Antonio", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119536158", "display_name": "Tareq Salem", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Salem, Tareq Si", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5048382270", "display_name": "Nicola Piovesan", "orcid": "https://orcid.org/0000-0001-5397-5248", "observed_orcids": ["https://orcid.org/0000-0001-5397-5248"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Piovesan, Nicola", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5064737739", "display_name": "Fadhel Ayed", "orcid": "https://orcid.org/0009-0008-4234-7682", "observed_orcids": ["https://orcid.org/0009-0008-4234-7682"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ayed, Fadhel", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "language": [1], "models": [2], "(LLMs)": [3], "rely": [4], "on": [5, 62], "key-value": [6], "(KV)": [7], "caches": [8], "for": [9, 147], "efficient": [10, 148], "autoregressive": [11], "decoding;": [12], "however,": [13], "cache": [14, 33, 58, 90], "size": [15], "grows": [16], "linearly": [17], "with": [18, 43, 112, 137], "context": [19], "length": [20], "and": [21, 80, 128, 144], "model": [22], "depth,": [23], "becoming": [24], "a": [25, 53, 142], "major": [26], "bottleneck": [27], "in": [28], "long-context": [29, 149], "inference.": [30], "Prior": [31], "KV": [32, 57], "compression": [34, 59], "methods": [35], "either": [36], "enforce": [37], "rigid": [38], "heuristics,": [39], "disrupt": [40], "tensor": [41], "layouts": [42], "per-attention-head": [44], "variability,": [45], "or": [46], "require": [47], "specialized": [48], "compute": [49], "kernels.": [50], "We": [51], "propose": [52], "simple,": [54], "yet": [55], "effective,": [56], "framework": [60], "based": [61], "attention-guided,": [63], "layer-adaptive": [64], "composite": [65, 84], "tokens.": [66, 114], "Our": [67], "method": [68], "aggregates": [69], "attention": [70], "scores": [71], "to": [72, 110], "estimate": [73], "token": [74], "importance,": [75], "selects": [76], "head-specific": [77], "tokens": [78, 85], "independently,": [79], "aligns": [81], "them": [82], "into": [83], "that": [86], "respect": [87], "the": [88], "uniform": [89], "structure": [91], "required": [92], "by": [93], "existing": [94], "inference": [95, 139], "engines.": [96], "A": [97], "global": [98], "allocation": [99], "mechanism": [100], "further": [101], "adapts": [102], "retention": [103], "budgets": [104], "across": [105], "layers,": [106], "assigning": [107], "more": [108], "capacity": [109], "layers": [111], "informative": [113], "This": [115], "approach": [116, 133], "achieves": [117], "significant": [118], "memory": [119], "reduction": [120], "while": [121], "preserving": [122], "accuracy,": [123], "consistently": [124], "outperforming": [125], "prior": [126], "structured": [127], "semi-structured": [129], "methods.": [130], "Crucially,": [131], "our": [132], "remains": [134], "fully": [135], "compatible": [136], "standard": [138], "pipelines,": [140], "offering": [141], "practical": [143], "scalable": [145], "solution": [146], "LLM": [150], "deployment.": [151]}, "primary_topic": {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.3935999870300293, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.3935999870300293, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.382099986076355, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.14100000262260437, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2509.05165", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2509.05165", "pdf_url": "https://arxiv.org/pdf/2509.05165", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7106338528", "doi": "https://doi.org/10.48550/arxiv.2511.15030", "title": "WiCo-PG: Wireless Channel Foundation Model for Pathloss Map Generation via Synesthesia of Machines", "publication_year": 2025, "publication_date": "2025-11-19", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Sun, Mingran", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sun, Mingran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Bai, Lu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bai, Lu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Huang, Ziwei", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Ziwei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Cai, Xuesong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cai, Xuesong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Cheng, Xiang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cheng, Xiang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Wu, Jianjun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Jianjun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"A": [0], "wireless": [1], "channel": [2], "foundation": [3], "model": [4, 145], "for": [5, 17, 37, 81, 112], "pathloss": [6, 61, 83, 114, 126], "map": [7, 62, 84, 115, 127], "generation": [8, 63, 85, 128], "(WiCo-PG)": [9], "via": [10], "Synesthesia": [11], "of": [12, 105, 139, 164], "Machines": [13], "(SoM)": [14], "is": [15, 35, 97, 110], "developed": [16], "the": [18, 53, 56, 121, 142, 151, 165, 171], "first": [19], "time.": [20], "Considering": [21], "sixth-generation": [22], "(6G)": [23], "uncrewed": [24], "aerial": [25], "vehicle": [26], "(UAV)-to-ground": [27], "(U2G)": [28], "scenarios,": [29, 43], "a": [30, 76, 100, 133], "new": [31], "multi-modal": [32], "sensing-communication": [33], "dataset": [34], "constructed": [36, 54], "WiCo-PG": [38, 58, 123, 167], "pre-training,": [39], "including": [40], "multiple": [41], "U2G": [42], "diverse": [44, 48], "flight": [45, 72], "altitudes,": [46], "and": [47, 71, 95, 150], "frequency": [49], "bands.": [50], "Based": [51], "on": [52, 87], "dataset,": [55], "proposed": [57, 122, 166], "enables": [59], "cross-modal": [60, 82, 113], "by": [64, 156, 173], "leveraging": [65], "RGB": [66], "images": [67], "from": [68], "different": [69], "scenarios": [70], "altitudes.": [73], "In": [74], "WiCo-PG,": [75], "novel": [77, 101], "network": [78], "architecture": [79, 109], "designed": [80, 111], "based": [86], "dual": [88], "vector": [89], "quantized": [90], "generative": [91], "adversarial": [92], "networks": [93], "(VQGANs)": [94], "Transformer": [96], "proposed.": [98], "Furthermore,": [99], "frequency-guided": [102], "shared-routed": [103], "mixture": [104], "experts": [106], "(S-R": [107], "MoE)": [108], "generation.": [116], "Simulation": [117], "results": [118], "demonstrate": [119], "that": [120], "achieves": [124], "improved": [125], "accuracy": [129], "through": [130], "pre-training": [131], "with": [132], "normalized": [134], "mean": [135], "squared": [136], "error": [137], "(NMSE)": [138], "0.012,": [140], "outperforming": [141], "large": [143], "language": [144], "(LLM)-based": [146], "scheme,": [147], "i.e.,": [148], "LLM4PG,": [149], "conventional": [152], "deep": [153], "learning-based": [154], "scheme": [155], "more": [157], "than": [158], "6.98": [159], "dB.": [160], "The": [161], "enhanced": [162], "generality": [163], "can": [168], "further": [169], "outperform": [170], "LLM4PG": [172], "at": [174], "least": [175], "1.37": [176], "dB": [177], "using": [178], "2.7%": [179], "samples": [180], "in": [181], "few-shot": [182], "generalization.": [183]}, "primary_topic": {"id": "https://openalex.org/T10936", "display_name": "Millimeter-Wave Propagation and Modeling", "score": 0.48069998621940613, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10936", "display_name": "Millimeter-Wave Propagation and Modeling", "score": 0.48069998621940613, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11133", "display_name": "UAV Applications and Optimization", "score": 0.2946999967098236, "subfield": {"id": "https://openalex.org/subfields/2202", "display_name": "Aerospace Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11458", "display_name": "Advanced Wireless Communication Technologies", "score": 0.08389999717473984, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2511.15030", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2511.15030", "pdf_url": "https://arxiv.org/pdf/2511.15030", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4406779184", "doi": "https://doi.org/10.48550/arxiv.2501.12766", "title": "NExtLong: Toward Effective Long-Context Training without Long Documents", "publication_year": 2025, "publication_date": "2025-01-22", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5025658951", "display_name": "Chaochen Gao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gao, Chaochen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083539852", "display_name": "Xing Wu", "orcid": "https://orcid.org/0000-0001-5331-022X", "observed_orcids": ["https://orcid.org/0000-0001-5331-022X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Xing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5108279234", "display_name": "Zijia Lin", "orcid": "https://orcid.org/0000-0001-7696-5368", "observed_orcids": ["https://orcid.org/0000-0001-7696-5368"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Zijia", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103145743", "display_name": "Debing Zhang", "orcid": "https://orcid.org/0000-0003-4048-0531", "observed_orcids": ["https://orcid.org/0000-0003-4048-0531"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Debing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5018288181", "display_name": "Songlin Hu", "orcid": "https://orcid.org/0000-0002-7170-3809", "observed_orcids": ["https://orcid.org/0000-0002-7170-3809"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hu, Songlin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "language": [1], "models": [2], "(LLMs)": [3], "with": [4], "extended": [5], "context": [6, 69, 88], "windows": [7], "have": [8], "made": [9], "significant": [10, 105], "strides": [11], "yet": [12], "remain": [13], "a": [14, 32, 48, 61], "challenge": [15], "due": [16], "to": [17, 26, 35, 84, 95, 115, 135], "the": [18, 37, 68, 82, 109], "scarcity": [19], "of": [20], "long": [21, 128, 140], "documents.": [22, 129], "Existing": [23], "methods": [24], "tend": [25], "synthesize": [27], "long-context": [28, 53, 117, 150], "data": [29, 54], "but": [30], "lack": [31], "clear": [33], "mechanism": [34], "reinforce": [36], "long-range": [38, 86, 97], "dependency": [39], "modeling.": [40], "To": [41], "address": [42], "this": [43], "limitation,": [44], "we": [45], "propose": [46], "NExtLong,": [47], "novel": [49], "framework": [50, 146], "for": [51, 147], "synthesizing": [52], "through": [55], "Negative": [56], "document": [57, 62], "Extension.": [58], "NExtLong": [59, 103], "decomposes": [60], "into": [63], "multiple": [64], "meta-chunks": [65], "and": [66, 111, 120], "extends": [67], "by": [70], "interleaving": [71], "hard": [72], "negative": [73], "distractors": [74], "retrieved": [75], "from": [76, 89], "pretraining": [77], "corpora.": [78], "This": [79], "approach": [80], "compels": [81], "model": [83, 96], "discriminate": [85], "dependent": [87], "distracting": [90], "content,": [91], "enhancing": [92], "its": [93], "ability": [94, 134], "dependencies.": [98], "Extensive": [99], "experiments": [100], "demonstrate": [101], "that": [102], "achieves": [104], "performance": [106], "improvements": [107], "on": [108, 126, 138], "HELMET": [110], "RULER": [112], "benchmarks": [113], "compared": [114], "existing": [116], "synthesis": [118], "approaches": [119], "leading": [121], "models,": [122], "which": [123], "are": [124], "trained": [125], "non-synthetic": [127, 139], "These": [130], "findings": [131], "highlight": [132], "NExtLong's": [133], "reduce": [136], "reliance": [137], "documents,": [141], "making": [142], "it": [143], "an": [144], "effective": [145], "developing": [148], "advanced": [149], "LLMs.": [151]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9901999831199646, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9901999831199646, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.005400000140070915, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.002099999925121665, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2501.12766", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2501.12766", "pdf_url": "https://arxiv.org/pdf/2501.12766", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "other-oa", "license_id": "https://openalex.org/licenses/other-oa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417295988", "doi": "https://doi.org/10.48550/arxiv.2512.10547", "title": "Unlocking the Address Book: Dissecting the Sparse Semantic Structure of LLM Key-Value Caches via Sparse Autoencoders", "publication_year": 2025, "publication_date": "2025-12-11", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5084704002", "display_name": "Qingsen Ma", "orcid": "https://orcid.org/0000-0001-8059-0458", "observed_orcids": ["https://orcid.org/0000-0001-8059-0458"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ma, Qingsen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wang, Dianyun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Dianyun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5046612778", "display_name": "Jiaming Lyu", "orcid": "https://orcid.org/0009-0003-8868-7248", "observed_orcids": ["https://orcid.org/0009-0003-8868-7248"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lyu, Jiaming", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wang, Yaoye", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yaoye", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5081265309", "display_name": "Ning Li", "orcid": "https://orcid.org/0000-0002-4534-9009", "observed_orcids": ["https://orcid.org/0000-0002-4534-9009", "https://orcid.org/0009-0005-1124-1672"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ning, Lechen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhu, Sujie", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Sujie", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100517463", "display_name": "Zhenbo Xu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Zhenbo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5077408167", "display_name": "Liuyu Xiang", "orcid": "https://orcid.org/0000-0001-8486-6255", "observed_orcids": ["https://orcid.org/0000-0001-8486-6255"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xiang, Liuyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5055511404", "display_name": "Huining Li", "orcid": "https://orcid.org/0000-0003-3865-3587", "observed_orcids": ["https://orcid.org/0000-0003-3865-3587"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Huining", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5019711261", "display_name": "Huijia Wu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Huijia", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5108370761", "display_name": "Zhaofeng He", "orcid": "https://orcid.org/0000-0003-3043-2122", "observed_orcids": ["https://orcid.org/0000-0003-3043-2122"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "He, Zhaofeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0], "Key-Value": [1], "(KV)": [2], "cache": [3, 42], "is": [4, 16], "the": [5, 40, 58, 108, 136, 141], "primary": [6], "memory": [7], "bottleneck": [8], "in": [9], "long-context": [10], "Large": [11], "Language": [12], "Models,": [13], "yet": [14], "it": [15], "typically": [17], "treated": [18], "as": [19, 76], "an": [20], "opaque": [21], "numerical": [22], "tensor.": [23], "In": [24], "this": [25, 98], "work,": [26], "we": [27, 100], "propose": [28], "\\textbf{STA-Attention},": [29], "a": [30, 68, 82, 93, 102], "framework": [31], "that": [32, 105, 125], "utilizes": [33], "Top-K": [34, 52], "Sparse": [35], "Autoencoders": [36], "(SAEs)": [37], "to": [38, 135], "decompose": [39], "KV": [41], "into": [43], "interpretable": [44], "``semantic": [45], "atoms.''": [46], "Unlike": [47], "standard": [48], "$L_1$-regularized": [49], "SAEs,": [50], "our": [51, 126], "approach": [53], "eliminates": [54], "shrinkage": [55], "bias,": [56], "preserving": [57], "precise": [59], "dot-product": [60], "geometry": [61], "required": [62], "for": [63], "attention.": [64], "Our": [65], "analysis": [66], "uncovers": [67], "fundamental": [69], "\\textbf{Key-Value": [70], "Asymmetry}:": [71], "while": [72, 113], "Key": [73], "vectors": [74, 87], "serve": [75], "highly": [77], "sparse": [78], "routers": [79], "dominated": [80], "by": [81], "``Semantic": [83], "Elbow,''": [84], "deep": [85], "Value": [86], "carry": [88], "dense": [89], "content": [90], "payloads": [91], "requiring": [92], "larger": [94], "budget.": [95], "Based": [96], "on": [97, 118], "structure,": [99], "introduce": [101], "Dual-Budget": [103], "Strategy": [104], "selectively": [106], "preserves": [107], "most": [109], "informative": [110], "semantic": [111, 127], "components": [112], "filtering": [114], "representational": [115], "noise.": [116], "Experiments": [117], "Yi-6B,": [119], "Mistral-7B,": [120], "Qwen2.5-32B,": [121], "and": [122, 131, 146], "others": [123], "show": [124], "reconstructions": [128], "maintain": [129], "perplexity": [130], "zero-shot": [132], "performance": [133], "comparable": [134], "original": [137], "models,": [138], "effectively": [139], "bridging": [140], "gap": [142], "between": [143], "mechanistic": [144], "interpretability": [145], "faithful": [147], "attention": [148], "modeling.": [149]}, "primary_topic": {"id": "https://openalex.org/T12026", "display_name": "Explainable Artificial Intelligence (XAI)", "score": 0.9589999914169312, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T12026", "display_name": "Explainable Artificial Intelligence (XAI)", "score": 0.9589999914169312, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.03539999946951866, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0012000000569969416, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.10547", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.10547", "pdf_url": "https://arxiv.org/pdf/2512.10547", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4413480762", "doi": "https://doi.org/10.3390/pr13092680", "title": "A Comparative Review of Large Language Models in Engineering with Emphasis on Chemical Engineering Applications", "publication_year": 2025, "publication_date": "2025-08-23", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Teck Leong Khoo", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405256456", "display_name": "Universiti Tunku Abdul Rahman Sungai Long Campus", "ror": "https://ror.org/046b54093", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}, {"id": "https://openalex.org/I931681460", "display_name": "Universiti Tunku Abdul Rahman", "ror": "https://ror.org/050pq4m56", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I931681460"]}], "countries": ["MY"], "is_corresponding": false, "raw_author_name": "Teck Leong Khoo", "raw_affiliation_strings": ["Department of Chemical Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Chemical Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia", "institution_ids": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Tin Sin Lee", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405256456", "display_name": "Universiti Tunku Abdul Rahman Sungai Long Campus", "ror": "https://ror.org/046b54093", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}, {"id": "https://openalex.org/I931681460", "display_name": "Universiti Tunku Abdul Rahman", "ror": "https://ror.org/050pq4m56", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I931681460"]}], "countries": ["MY"], "is_corresponding": true, "raw_author_name": "Tin Sin Lee", "raw_affiliation_strings": ["Department of Chemical Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Chemical Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia", "institution_ids": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5069258975", "display_name": "Soo\u2010Tueen Bee", "orcid": "https://orcid.org/0000-0002-3578-8737", "observed_orcids": ["https://orcid.org/0000-0002-3578-8737"]}, "institutions": [{"id": "https://openalex.org/I4405256456", "display_name": "Universiti Tunku Abdul Rahman Sungai Long Campus", "ror": "https://ror.org/046b54093", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}, {"id": "https://openalex.org/I931681460", "display_name": "Universiti Tunku Abdul Rahman", "ror": "https://ror.org/050pq4m56", "country_code": "MY", "type": "education", "lineage": ["https://openalex.org/I931681460"]}], "countries": ["MY"], "is_corresponding": true, "raw_author_name": "Soo-Tueen Bee", "raw_affiliation_strings": ["Department of Mechanical and Material Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Mechanical and Material Engineering, Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Bandar Sungai Long, Cheras, Kajang 43000, Selangor, Malaysia", "institution_ids": ["https://openalex.org/I4405256456", "https://openalex.org/I931681460"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5053918728", "display_name": "Chi Ma", "orcid": "https://orcid.org/0000-0002-5127-0487", "observed_orcids": ["https://orcid.org/0000-0002-5127-0487"]}, "institutions": [{"id": "https://openalex.org/I48780066", "display_name": "Shenyang University of Chemical Technology", "ror": "https://ror.org/03dbpdh75", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I48780066"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chi Ma", "raw_affiliation_strings": ["School of Materials Science and Engineering, Shenyang University of Chemical Technology, Shenyang Economical and Technological Development Zone, Street No. 11, Shenyang 110142, China"], "raw_orcid": "https://orcid.org/0000-0002-5127-0487", "affiliations": [{"raw_affiliation_string": "School of Materials Science and Engineering, Shenyang University of Chemical Technology, Shenyang Economical and Technological Development Zone, Street No. 11, Shenyang 110142, China", "institution_ids": ["https://openalex.org/I48780066"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5100320635", "display_name": "Yuanyuan Zhang", "orcid": "https://orcid.org/0000-0001-9738-604X", "observed_orcids": ["https://orcid.org/0000-0001-9738-604X"]}, "institutions": [{"id": "https://openalex.org/I4405278512", "display_name": "Shandong Institute of Petroleum and Chemical Technology", "ror": "https://ror.org/00581xj59", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I4405278512"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yuan-Yuan Zhang", "raw_affiliation_strings": ["School of Chemical Engineering, Shandong Institute of Petroleum and Chemical Technology, No. 1 Ji\u2019nan Road, Dongying District, Dongying 257000, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Chemical Engineering, Shandong Institute of Petroleum and Chemical Technology, No. 1 Ji\u2019nan Road, Dongying District, Dongying 257000, China", "institution_ids": ["https://openalex.org/I4405278512"]}]}], "abstract_inverted_index": {"This": [0], "review": [1, 30, 108, 162, 200], "provides": [2], "a": [3, 23], "comprehensive": [4], "overview": [5], "of": [6, 11, 35, 50, 61, 115, 141, 191, 207], "the": [7, 32, 47, 58, 66, 113, 138, 189, 199, 202, 214], "evolution": [8], "and": [9, 15, 41, 53, 65, 74, 121, 130, 154, 158, 171, 183, 204, 221], "application": [10], "artificial": [12], "intelligence": [13], "(AI)": [14], "large": [16], "language": [17, 72], "models": [18, 43, 62], "(LLMs)": [19], "in": [20, 69, 92, 117, 143, 179], "engineering,": [21, 123, 145], "with": [22, 101], "specific": [24, 165], "focus": [25], "on": [26, 112, 127, 168], "chemical": [27, 144], "engineering.": [28], "The": [29, 107, 161], "traces": [31], "historical": [33], "development": [34], "LLMs,": [36], "from": [37], "early": [38], "rule-based": [39], "systems": [40], "statistical": [42], "like": [44, 63], "N-grams": [45], "to": [46, 89, 137], "transformative": [48], "introduction": [49], "neural": [51], "networks": [52], "transformer": [54], "architecture.": [55], "It": [56], "examines": [57], "pivotal": [59], "role": [60], "BERT": [64], "GPT": [67], "series": [68], "advancing": [70], "natural": [71], "processing": [73], "enabling": [75], "sophisticated": [76], "applications": [77, 140], "across": [78], "various": [79], "engineering": [80, 211], "disciplines.": [81], "For": [82], "example,": [83], "GPT-3": [84], "(175B": [85], "parameters)": [86], "demonstrates": [87], "up": [88], "87.7%": [90], "accuracy": [91], "structured": [93], "information": [94], "extraction,": [95], "while": [96, 186], "GPT-4": [97], "introduces": [98], "multimodal": [99], "reasoning": [100], "estimated": [102], "token": [103], "limits": [104], "exceeding": [105], "32k.": [106], "synthesizes": [109], "recent": [110], "research": [111], "use": [114, 148], "LLMs": [116, 142, 176, 209], "software,": [118], "mechanical,": [119], "civil,": [120], "electrical": [122], "highlighting": [124], "their": [125, 147, 193], "impact": [126], "automation,": [128], "design,": [129, 173], "decision-making.": [131], "A": [132], "significant": [133], "portion": [134], "is": [135], "dedicated": [136], "burgeoning": [139], "including": [146], "as": [149], "educational": [150], "tools,": [151], "process": [152], "simulation": [153], "modelling,": [155], "reaction": [156], "optimization,": [157], "molecular": [159], "design.": [160], "delves": [163], "into": [164, 210], "case": [166], "studies": [167], "distillation": [169], "column": [170], "reactor": [172], "showcasing": [174], "how": [175], "can": [177], "assist": [178], "generating": [180], "initial": [181], "parameters": [182], "optimizing": [184], "processes": [185], "also": [187], "underscoring": [188], "necessity": [190], "validating": [192], "outputs": [194], "against": [195], "traditional": [196], "methods.": [197], "Finally,": [198], "addresses": [201], "challenges": [203], "future": [205], "considerations": [206], "integrating": [208], "workflows,": [212], "emphasizing": [213], "need": [215], "for": [216], "domain-specific": [217], "adaptations,": [218], "ethical": [219], "guidelines,": [220], "robust": [222], "validation": [223], "frameworks.": [224]}, "primary_topic": {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.8672000169754028, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.8672000169754028, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.08129999786615372, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11948", "display_name": "Machine Learning in Materials Science", "score": 0.020099999383091927, "subfield": {"id": "https://openalex.org/subfields/2505", "display_name": "Materials Chemistry"}, "field": {"id": "https://openalex.org/fields/25", "display_name": "Materials Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.3390/pr13092680", "is_oa": true, "landing_page_url": "https://doi.org/10.3390/pr13092680", "pdf_url": "https://www.mdpi.com/2227-9717/13/9/2680/pdf?version=1755932248", "source": {"id": "https://openalex.org/S4210201879", "display_name": "Processes", "issn_l": "2227-9717", "issn": ["2227-9717"], "is_oa": true, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core"], "host_organization": "https://openalex.org/P4310310987", "host_organization_name": "Multidisciplinary Digital Publishing Institute", "host_organization_lineage": ["https://openalex.org/P4310310987"], "host_organization_lineage_names": ["Multidisciplinary Digital Publishing Institute"], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Processes", "raw_type": "journal-article"}, "cited_by_count": 15, "is_retracted": false}, {"id": "https://openalex.org/W4407208894", "doi": "https://doi.org/10.48550/arxiv.2502.02659", "title": "A Training-Free Length Extrapolation Approach for LLMs: Greedy Attention Logit Interpolation (GALI)", "publication_year": 2025, "publication_date": "2025-02-04", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5101930357", "display_name": "Yan Li", "orcid": "https://orcid.org/0000-0003-3222-1257", "observed_orcids": ["https://orcid.org/0000-0003-3222-1257"]}, "institutions": [{"id": "https://openalex.org/I129604602", "display_name": "The University of Sydney", "ror": "https://ror.org/0384j8v12", "country_code": "AU", "type": "education", "lineage": ["https://openalex.org/I129604602"]}], "countries": ["AU"], "is_corresponding": false, "raw_author_name": "Li, Yan", "raw_affiliation_strings": ["The University of Sydney ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The University of Sydney ,", "institution_ids": ["https://openalex.org/I129604602"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100437442", "display_name": "Tianyi Zhang", "orcid": "https://orcid.org/0000-0001-6956-5622", "observed_orcids": ["https://orcid.org/0000-0001-6956-5622", "https://orcid.org/0009-0001-8703-0100"]}, "institutions": [{"id": "https://openalex.org/I165779595", "display_name": "The University of Melbourne", "ror": "https://ror.org/01ej9dk98", "country_code": "AU", "type": "education", "lineage": ["https://openalex.org/I165779595"]}], "countries": ["AU"], "is_corresponding": false, "raw_author_name": "Zhang, Tianyi", "raw_affiliation_strings": ["The University of Melbourne ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The University of Melbourne ,", "institution_ids": ["https://openalex.org/I165779595"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100625753", "display_name": "Zhi Li", "orcid": "https://orcid.org/0000-0003-0944-9886", "observed_orcids": ["https://orcid.org/0000-0003-0944-9886"]}, "institutions": [{"id": "https://openalex.org/I16609230", "display_name": "Hunan University", "ror": "https://ror.org/05htk5m33", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I16609230"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Li, Zechuan", "raw_affiliation_strings": ["Hunan University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Hunan University", "institution_ids": ["https://openalex.org/I16609230"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5084419965", "display_name": "Soyeon Caren Han", "orcid": "https://orcid.org/0000-0002-1948-6819", "observed_orcids": ["https://orcid.org/0000-0002-1948-6819"]}, "institutions": [{"id": "https://openalex.org/I165779595", "display_name": "The University of Melbourne", "ror": "https://ror.org/01ej9dk98", "country_code": "AU", "type": "education", "lineage": ["https://openalex.org/I165779595"]}], "countries": ["AU"], "is_corresponding": false, "raw_author_name": "Han, Soyeon Caren", "raw_affiliation_strings": ["The University of Melbourne ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The University of Melbourne ,", "institution_ids": ["https://openalex.org/I165779595"]}]}], "abstract_inverted_index": {"Transformer-based": [0], "Large": [1], "Language": [2], "Models": [3], "(LLMs)": [4], "struggle": [5], "with": [6, 128], "inputs": [7], "exceeding": [8], "their": [9], "training": [10], "context": [11], "window": [12], "due": [13], "to": [14, 67, 101], "positional": [15, 41, 61, 94], "out-of-distribution": [16], "(O.O.D.)": [17], "issues": [18], "that": [19, 53, 91, 98], "disrupt": [20], "attention.": [21], "Existing": [22], "solutions,": [23], "including": [24], "fine-tuning": [25], "and": [26, 63, 73, 97, 117], "training-free": [27, 51], "methods,": [28], "face": [29], "challenges": [30], "like": [31], "inefficiency,": [32], "redundant": [33], "interpolation,": [34], "logit": [35, 66], "outliers,": [36], "or": [37], "loss": [38], "of": [39, 80, 125], "local": [40], "information.": [42], "We": [43], "propose": [44], "Greedy": [45], "Attention": [46], "Logit": [47], "Interpolation": [48], "(GALI),": [49], "a": [50, 77, 112], "method": [52], "improves": [54, 104], "length": [55], "extrapolation": [56], "by": [57], "greedily": [58], "reusing": [59], "pretrained": [60], "intervals": [62, 95], "interpolating": [64], "attention": [65], "eliminate": [68], "outliers.": [69], "GALI": [70, 110], "achieves": [71], "stable": [72], "superior": [74], "performance": [75], "across": [76], "wide": [78], "range": [79], "long-context": [81], "tasks": [82], "without": [83], "requiring": [84], "input-length-specific": [85], "tuning.": [86], "Our": [87, 123], "analysis": [88], "further": [89], "reveals": [90], "LLMs": [92], "interpret": [93], "unevenly": [96], "restricting": [99], "interpolation": [100], "narrower": [102], "ranges": [103], "performance,": [105], "even": [106], "on": [107], "short-context": [108], "tasks.": [109], "represents": [111], "step": [113], "toward": [114], "more": [115], "robust": [116], "generalizable": [118], "long-text": [119], "processing": [120], "in": [121], "LLMs.": [122], "implementation": [124], "GALI,": [126], "along": [127], "the": [129], "experiments": [130], "from": [131], "our": [132], "paper,": [133], "is": [134], "open-sourced": [135], "at": [136], "https://github.com/adlnlp/Gali.": [137]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9628999829292297, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9628999829292297, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.0272000003606081, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.003800000064074993, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.02659", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.02659", "pdf_url": "https://arxiv.org/pdf/2502.02659", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417092559", "doi": "https://doi.org/10.48550/arxiv.2505.10570", "title": "LongFuncEval: Measuring the effectiveness of long context models for function calling", "publication_year": 2025, "publication_date": "2025-04-30", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5082592545", "display_name": "Kiran Kate", "orcid": "https://orcid.org/0009-0003-9688-9245", "observed_orcids": ["https://orcid.org/0009-0003-9688-9245"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kate, Kiran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066128057", "display_name": "Tejaswini Pedapati", "orcid": "https://orcid.org/0000-0002-5260-0951", "observed_orcids": ["https://orcid.org/0000-0002-5260-0951"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Pedapati, Tejaswini", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5053808698", "display_name": "Kinjal Basu", "orcid": "https://orcid.org/0000-0001-8693-9307", "observed_orcids": ["https://orcid.org/0000-0001-8693-9307"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Basu, Kinjal", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5058002550", "display_name": "Yara Rizk", "orcid": "https://orcid.org/0000-0001-7475-2385", "observed_orcids": ["https://orcid.org/0000-0001-7475-2385"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Rizk, Yara", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5058823844", "display_name": "Vijil Chenthamarakshan", "orcid": "https://orcid.org/0000-0001-7830-5777", "observed_orcids": ["https://orcid.org/0000-0001-7830-5777"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chenthamarakshan, Vijil", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5000750466", "display_name": "Subhajit Chaudhury", "orcid": "https://orcid.org/0000-0003-3435-2584", "observed_orcids": ["https://orcid.org/0000-0003-3435-2584"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chaudhury, Subhajit", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5076160843", "display_name": "Mayank Agarwal", "orcid": "https://orcid.org/0000-0002-6374-6737", "observed_orcids": ["https://orcid.org/0000-0002-6374-6737", "https://orcid.org/0009-0004-0582-2836"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Agarwal, Mayank", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5031546123", "display_name": "Ibrahim Abdelaziz", "orcid": "https://orcid.org/0000-0003-1449-5115", "observed_orcids": ["https://orcid.org/0000-0003-1449-5115"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Abdelaziz, Ibrahim", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Multiple": [0], "recent": [1], "studies": [2], "have": [3], "documented": [4], "large": [5, 58, 98], "language": [6, 125], "models'": [7], "(LLMs)": [8], "performance": [9, 201], "on": [10, 16], "calling": [11, 47, 149, 250], "external": [12], "tools/functions.": [13], "Others": [14], "focused": [15], "LLMs'": [17, 34], "abilities": [18, 35], "to": [19, 36, 79, 91, 135, 169, 205, 215, 255], "handle": [20], "longer": [21], "context": [22, 43, 110, 140, 178, 196, 247], "lengths.": [23], "At": [24], "the": [25, 66, 116, 121, 132, 138, 147, 176, 183, 186, 189, 208, 222], "intersection": [26], "of": [27, 60, 82, 100, 108, 143, 185, 203, 210], "these": [28, 144], "areas": [29], "lies": [30], "another": [31], "interesting": [32], "problem:": [33], "accurately": [37], "perform": [38], "function": [39], "calls": [40], "in": [41, 86, 115, 120, 146, 188, 218, 248], "long": [42, 63, 71, 109, 139, 195, 246], "settings.": [44], "Particularly,": [45], "when": [46], "tools,": [48, 61], "LLMs": [49, 83, 242], "are": [50, 76], "encumbered": [51], "by": [52], "three": [53], "predominant": [54], "challenges:": [55], "(1)": [56], "a": [57, 97, 162, 200, 213], "catalog": [59, 99], "(2)": [62], "responses": [64, 224], "from": [65], "tool": [67, 148, 223, 249], "APIs,": [68], "and": [69, 158, 160, 180, 227, 229], "(3)": [70], "multi-turn": [72, 87, 234], "conversations.": [73], "These": [74], "challenges": [75, 111], "particularly": [77], "relevant": [78], "enterprise": [80], "applications": [81], "which": [84], "engage": [85], "conversations": [88, 235], "with": [89, 193, 245], "users": [90], "complete": [92], "complex": [93, 101], "tasks": [94], "that": [95, 241], "require": [96], "tools.": [102], "The": [103], "literature": [104], "contains": [105], "multiple": [106], "investigations": [107], "such": [112], "as": [113, 207, 221, 233], "lost": [114], "middle": [117], "or": [118], "needle": [119], "haystack": [122], "for": [123, 155, 166, 232], "natural": [124], "tasks.": [126], "In": [127], "this": [128, 171], "paper,": [129], "we": [130, 198], "make": [131], "first": [133], "attempt": [134], "comprehensively": [136], "study": [137, 239], "understanding": [141], "capabilities": [142], "models": [145], "setup.": [150], "We": [151, 173], "modify": [152], "existing": [153], "benchmarks": [154], "challenge": [156, 167], "1": [157], "3,": [159], "create": [161], "new": [163], "evaluation": [164], "set": [165], "2": [168], "enable": [170], "analysis.": [172], "gradually": [174], "increase": [175], "input": [177], "length": [179, 225], "also": [181], "vary": [182], "position": [184], "answer": [187, 219], "input.": [190], "When": [191], "evaluated": [192], "several": [194], "models,": [197], "observe": [199], "drop": [202], "7%": [204, 214], "85%": [206], "number": [209], "tools": [211], "increases,": [212, 226], "91%": [216], "degradation": [217, 231], "retrieval": [220], "13%": [228], "40%": [230], "get": [236], "longer.": [237], "Our": [238], "shows": [240], "still": [243], "struggle": [244], "settings,": [251], "motivating": [252], "future": [253], "research": [254], "drive": [256], "further": [257], "LLM": [258], "improvements.": [259]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7035999894142151, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7035999894142151, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.10209999978542328, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.06759999692440033, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2505.10570", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2505.10570", "pdf_url": "https://arxiv.org/pdf/2505.10570", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4407953097", "doi": "https://doi.org/10.31219/osf.io/9ds4a_v1", "title": "Hands-On with Transformers: A Step-by-Step Guide to Building Language Models from Scratch", "publication_year": 2025, "publication_date": "2025-02-26", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5116772880", "display_name": "Mo Ahtasam", "orcid": "https://orcid.org/0009-0000-3276-7335", "observed_orcids": ["https://orcid.org/0009-0000-3276-7335"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "MO AHTASAM", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0000-3276-7335", "affiliations": []}], "abstract_inverted_index": {"The": [0], "advent": [1], "of": [2, 75, 108], "the": [3, 39, 72, 105, 131], "Transformer": [4], "architecture": [5], "has": [6], "revolutionized": [7], "Natural": [8], "Language": [9], "Processing": [10], "(NLP),": [11], "rendering": [12], "traditional": [13], "recurrent": [14], "neural": [15], "networks": [16], "(RNNs)": [17], "obsolete": [18], "through": [19, 58], "innovations": [20], "like": [21], "self-attention": [22], "and": [23, 49, 80, 127, 136], "parallelization.": [24], "This": [25], "review": [26], "paper": [27, 103], "serves": [28], "a": [29, 35, 59], "dual": [30], "purpose:": [31], "(1)": [32], "to": [33, 53, 82, 120], "provide": [34], "theoretical": [36, 134], "deep-dive": [37], "into": [38, 67], "Transformer\u2019s": [40], "architecture,": [41], "dissecting": [42], "its": [43], "core": [44], "components\u2014self-attention": [45], "mechanisms,": [46], "positional": [47], "encoding,": [48], "encoder-decoder": [50], "frameworks\u2014and": [51], "(2)": [52], "deliver": [54], "practical,": [55], "educational": [56], "value": [57], "modular,": [60], "open-source": [61], "PyTorch": [62], "implementation.": [63], "By": [64, 129], "translating": [65], "theory": [66], "executable": [68], "code,": [69], "we": [70], "demystify": [71], "\"black": [73], "box\"": [74], "Transformers,": [76], "enabling": [77], "researchers,": [78], "educators,": [79], "developers": [81], "experiment": [83], "with": [84], "state-of-the-art": [85], "NLP": [86], "tools": [87], "even": [88], "under": [89], "hardware": [90], "constraints.Our": [91], "implementation,": [92], "tested": [93], "on": [94], "an": [95], "English-Hindi": [96], "translation": [97], "task.": [98], "Beyond": [99], "technical": [100], "analysis,": [101], "this": [102], "emphasizes": [104], "social": [106], "impact": [107], "democratizing": [109], "AI": [110], "education,": [111], "showing": [112], "how": [113], "accessible": [114], "frameworks": [115], "can": [116], "empower": [117], "global": [118], "communities": [119], "address": [121], "challenges": [122], "in": [123], "language": [124], "preservation,": [125], "healthcare,": [126], "education.": [128], "bridging": [130], "gap": [132], "between": [133], "understanding": [135], "hands-on": [137], "application.": [138]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7182999849319458, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7182999849319458, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.2736999988555908, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.001500000013038516, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.31219/osf.io/9ds4a_v1", "is_oa": true, "landing_page_url": "https://doi.org/10.31219/osf.io/9ds4a_v1", "pdf_url": null, "source": null, "license": "public-domain", "license_id": "https://openalex.org/licenses/public-domain", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415877255", "doi": "https://doi.org/10.48550/arxiv.2510.15231", "title": "Extending Audio Context for Long-Form Understanding in Large Audio-Language Models", "publication_year": 2025, "publication_date": "2025-10-17", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5120249624", "display_name": "Yuatyong Chaichana", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chaichana, Yuatyong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5088392357", "display_name": "Pittawat Taveekitworachai", "orcid": "https://orcid.org/0000-0002-6824-2634", "observed_orcids": ["https://orcid.org/0000-0002-6824-2634"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Taveekitworachai, Pittawat", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5114387463", "display_name": "Warit Sirichotedumrong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sirichotedumrong, Warit", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5024447363", "display_name": "Potsawee Manakul", "orcid": "https://orcid.org/0000-0001-7108-8626", "observed_orcids": ["https://orcid.org/0000-0001-7108-8626"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Manakul, Potsawee", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5021237405", "display_name": "Kunat Pipatanakul", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Pipatanakul, Kunat", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "Audio-Language": [1], "Models": [2], "(LALMs)": [3], "are": [4], "often": [5], "constrained": [6], "by": [7], "short": [8], "audio": [9, 22, 60, 97, 139], "context": [10, 46], "windows,": [11], "even": [12], "when": [13], "their": [14, 36], "text": [15, 64, 72], "backbones": [16], "support": [17], "long": [18, 138], "contexts,": [19], "limiting": [20], "long-form": [21], "understanding.": [23], "Prior": [24], "work": [25], "has": [26], "introduced": [27], "context-extension": [28], "methods": [29], "(e.g.": [30], "YaRN)": [31], "on": [32, 44, 114, 137], "unimodal": [33], "LLMs,": [34], "yet": [35], "application": [37], "to": [38, 67, 103], "LALMs": [39], "remains": [40], "unexplored.": [41], "First,": [42], "building": [43], "RoPE-based": [45], "extension,": [47], "we": [48, 75], "introduce": [49], "Partial": [50, 87, 120], "YaRN,": [51], "a": [52, 82, 90], "training-free,": [53], "modality-decoupled": [54], "extension": [55], "method": [56], "that": [57, 85, 119], "modifies": [58], "only": [59], "token": [61], "positions,": [62], "leaving": [63], "positions": [65], "intact": [66], "preserve": [68], "the": [69, 123], "base": [70], "LLM's": [71], "capabilities.": [73], "Second,": [74], "propose": [76], "Virtual": [77], "Longform": [78], "Audio": [79], "Training": [80], "(VLAT),": [81], "training": [83], "strategy": [84], "extends": [86], "YaRN": [88, 121], "into": [89], "training-time": [91], "positional": [92], "augmentation.": [93], "VLAT": [94, 132], "simulates": [95], "diverse": [96], "lengths": [98], "during": [99], "training,": [100], "enabling": [101], "generalization": [102], "inputs": [104], "far": [105], "longer": [106], "than": [107], "those": [108], "seen": [109], "in": [110], "training.": [111], "Our": [112], "experiments": [113], "SALMONN": [115], "and": [116, 131], "Qwen2-Audio": [117], "confirm": [118], "outperforms": [122], "original": [124], "models": [125], "across": [126], "wide": [127], "range": [128], "of": [129, 140], "settings,": [130], "provides": [133], "substantial": [134], "performance": [135], "improvement": [136], "unseen": [141], "lengths.": [142]}, "primary_topic": {"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.6176999807357788, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.6176999807357788, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.25459998846054077, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11309", "display_name": "Music and Audio Processing", "score": 0.05660000070929527, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.15231", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2510.15231", "pdf_url": "https://arxiv.org/pdf/2510.15231", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416447653", "doi": "https://doi.org/10.48550/arxiv.2505.16056", "title": "Not All Models Suit Expert Offloading: On Local Routing Consistency of Mixture-of-Expert Models", "publication_year": 2025, "publication_date": "2025-05-21", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100766214", "display_name": "Jingcong Liang", "orcid": "https://orcid.org/0009-0000-2488-2683", "observed_orcids": ["https://orcid.org/0009-0000-2488-2683"]}, "institutions": [{"id": "https://openalex.org/I24943067", "display_name": "Fudan University", "ror": "https://ror.org/013q1eq08", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I24943067"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Liang, Jingcong", "raw_affiliation_strings": ["Fudan University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Fudan University", "institution_ids": ["https://openalex.org/I24943067"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100440540", "display_name": "Siyuan Wang", "orcid": "https://orcid.org/0000-0001-7174-9754", "observed_orcids": ["https://orcid.org/0000-0001-7174-9754", "https://orcid.org/0009-0003-7428-8609"]}, "institutions": [{"id": "https://openalex.org/I1174212", "display_name": "University of Southern California", "ror": "https://ror.org/03taz7m60", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1174212"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Wang, Siyuan", "raw_affiliation_strings": ["University of Southern California"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Southern California", "institution_ids": ["https://openalex.org/I1174212"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049254205", "display_name": "Miren Tian", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tian, Miren", "raw_affiliation_strings": ["Huawei Technologies Ltd"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Huawei Technologies Ltd", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100413875", "display_name": "Yitong Li", "orcid": "https://orcid.org/0000-0002-6753-6227", "observed_orcids": ["https://orcid.org/0000-0002-6753-6227", "https://orcid.org/0009-0002-9020-8026"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Yitong", "raw_affiliation_strings": ["Huawei Technologies Ltd"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Huawei Technologies Ltd", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110628758", "display_name": "Duyu Tang", "orcid": "https://orcid.org/0009-0005-0667-6583", "observed_orcids": ["https://orcid.org/0009-0005-0667-6583"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Duyu", "raw_affiliation_strings": ["Huawei Technologies Ltd"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Huawei Technologies Ltd", "institution_ids": []}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5011504177", "display_name": "Zhongyu Wei", "orcid": "https://orcid.org/0000-0003-3789-8507", "observed_orcids": ["https://orcid.org/0000-0003-3789-8507"]}, "institutions": [{"id": "https://openalex.org/I24943067", "display_name": "Fudan University", "ror": "https://ror.org/013q1eq08", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I24943067"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wei, Zhongyu", "raw_affiliation_strings": ["Fudan University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Fudan University", "institution_ids": ["https://openalex.org/I24943067"]}]}], "abstract_inverted_index": {"Mixture-of-Experts": [0], "(MoE)": [1], "enables": [2], "efficient": [3], "scaling": [4], "of": [5, 34, 59, 70, 93, 109, 115, 118, 133, 140], "large": [6, 19], "language": [7], "models": [8, 21, 77, 160, 229], "(LLMs)": [9], "with": [10, 152, 191, 236], "sparsely": [11], "activated": [12], "experts": [13, 35, 110, 199, 217], "during": [14], "inference.": [15], "To": [16], "effectively": [17], "deploy": [18], "MoE": [20, 94, 150, 251], "on": [22, 41, 46, 50], "memory-constrained": [23], "devices,": [24], "many": [25], "systems": [26], "introduce": [27], "*expert": [28], "offloading*": [29], "that": [30, 185, 200, 215, 227], "caches": [31], "a": [32, 106, 116, 138, 144, 172], "subset": [33], "in": [36], "fast": [37], "memory,": [38], "leaving": [39], "others": [40], "slow": [42], "memory": [43], "to": [44, 88, 161, 166, 207, 220], "run": [45], "CPU": [47], "or": [48], "load": [49, 181, 187], "demand.": [51], "While": [52], "some": [53], "research": [54], "has": [55], "exploited": [56], "the": [57, 68, 113, 130, 241, 247, 261], "locality": [58], "expert": [60, 135, 202], "activations,": [61], "where": [62], "consecutive": [63], "tokens": [64], "activate": [65], "similar": [66], "experts,": [67], "degree": [69], "this": [71, 82], "**local": [72], "routing": [73, 91, 168, 177, 193, 210, 221], "consistency**": [74], "varies": [75], "across": [76], "and": [78, 120, 155, 157, 179, 226, 234, 253], "remains": [79], "understudied.": [80], "In": [81], "paper,": [83], "we": [84], "propose": [85], "two": [86], "metrics": [87], "measure": [89], "local": [90, 167, 176, 192, 209], "consistency": [92, 178, 222], "models:": [95], "(1)": [96], "**Segment": [97, 122], "Routing": [98], "Best": [99, 124], "Performance": [100], "(SRP)**,": [101], "which": [102, 128], "evaluates": [103], "how": [104], "well": [105], "fixed": [107], "group": [108], "can": [111, 189, 205], "cover": [112], "needs": [114], "segment": [117], "tokens,": [119], "(2)": [121], "Cache": [123], "Hit": [125], "Rate": [126], "(SCH)**,": [127], "measures": [129], "hit": [131], "rate": [132], "an": [134], "cache": [136, 145, 232, 237], "utilizing": [137], "length": [139], "future": [141], "information": [142], "under": [143], "limit.": [146], "We": [147, 170, 212, 259], "analyze": [148], "20": [149], "LLMs": [151], "diverse": [153], "sizes": [154, 238], "architectures": [156], "use": [158], "toy": [159], "verify": [162], "key": [163], "factors": [164], "related": [165], "consistency.": [169, 194, 211], "find": [171], "strong": [173], "trade-off": [174], "between": [175, 231], "*local*": [180], "balance,": [182], "while": [183], "showing": [184], "*global*": [186], "balance": [188, 230], "coexist": [190], "Meanwhile,": [195], "settings": [196], "like": [197], "shared": [198], "decrease": [201], "combination": [203], "space": [204], "lead": [206], "low": [208], "further": [213], "reveal": [214], "domain-specialized": [216], "contribute": [218], "more": [219], "than": [223], "vocabulary-specialized": [224], "ones,": [225], "most": [228], "effectiveness": [233], "efficiency": [235], "approximately": [239], "twice": [240], "active": [242], "experts.": [243], "These": [244], "findings": [245], "pave": [246], "way": [248], "for": [249, 263], "memory-efficient": [250], "design": [252], "deployment": [254], "without": [255], "compromising": [256], "inference": [257], "speed.": [258], "publish": [260], "code": [262], "replicating": [264], "experiments": [265], "at": [266], "https://github.com/ljcleo/moe-lrc": [267], ".": [268]}, "primary_topic": {"id": "https://openalex.org/T10054", "display_name": "Parallel Computing and Optimization Techniques", "score": 0.5932999849319458, "subfield": {"id": "https://openalex.org/subfields/1708", "display_name": "Hardware and Architecture"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10054", "display_name": "Parallel Computing and Optimization Techniques", "score": 0.5932999849319458, "subfield": {"id": "https://openalex.org/subfields/1708", "display_name": "Hardware and Architecture"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.28200000524520874, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.042399998754262924, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2505.16056", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2505.16056", "pdf_url": "https://arxiv.org/pdf/2505.16056", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416949783", "doi": "https://doi.org/10.48550/arxiv.2511.22955", "title": "Experts are all you need: A Composable Framework for Large Language Model Inference", "publication_year": 2025, "publication_date": "2025-11-28", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5028830082", "display_name": "Shrihari Sridharan", "orcid": "https://orcid.org/0000-0003-2434-4767", "observed_orcids": ["https://orcid.org/0000-0003-2434-4767"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sridharan, Shrihari", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083686495", "display_name": "Sourjya Roy", "orcid": "https://orcid.org/0000-0002-6864-7738", "observed_orcids": ["https://orcid.org/0000-0002-6864-7738"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Roy, Sourjya", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5065766721", "display_name": "Anand Raghunathan", "orcid": "https://orcid.org/0000-0002-4624-564X", "observed_orcids": ["https://orcid.org/0000-0002-4624-564X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Raghunathan, Anand", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5031161187", "display_name": "Kaushik Roy", "orcid": "https://orcid.org/0000-0002-0735-9695", "observed_orcids": ["https://orcid.org/0000-0002-0735-9695"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Roy, Kaushik", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs)": [3], "have": [4], "achieved": [5], "state-of-the-art": [6], "accuracies": [7], "in": [8, 158, 207, 219], "a": [9, 50, 108, 148, 183], "variety": [10], "of": [11, 24, 35, 52, 62, 124, 200], "natural": [12], "language": [13], "processing": [14], "(NLP)": [15], "tasks.": [16], "However,": [17, 56, 88], "this": [18, 39], "success": [19], "comes": [20], "at": [21], "the": [22, 67, 159, 216], "cost": [23], "increased": [25], "model": [26, 43, 72, 208, 218], "sizes": [27], "which": [28, 96], "leads": [29], "to": [30, 139, 193, 215, 229], "additional": [31], "computational": [32], "burden.": [33], "Mixture": [34], "Experts": [36], "(MoEs)": [37], "overcome": [38], "bottleneck": [40], "by": [41, 47, 81, 106], "decoupling": [42], "capacity": [44], "from": [45], "computation": [46], "only": [48], "activating": [49], "subset": [51], "parameters": [53], "or": [54], "\"experts\".": [55], "these": [57, 63, 89, 104], "models": [58], "require": [59], "joint": [60], "pretraining": [61], "experts": [64], "along": [65], "with": [66, 210], "router": [68], "and": [69, 146, 161, 169, 172], "do": [70], "not": [71], "multi-step": [73], "reasoning.": [74], "In": [75], "contrast,": [76], "multi-agent": [77], "frameworks": [78, 90], "improve": [79], "reasoning": [80], "decomposing": [82], "complex": [83], "problems": [84], "into": [85, 182], "modular": [86], "subtasks.": [87], "rely": [91], "on": [92, 167], "sequential": [93, 230], "\"plan--act--observe\"": [94], "loops,": [95], "introduce": [97], "significant": [98, 212], "latency.": [99], "Our": [100], "work,": [101], "Comp-LLM,": [102], "addresses": [103], "challenges": [105], "introducing": [107], "composable": [109], "inference": [110], "framework": [111], "that": [112, 131, 155, 177], "enables": [113], "cross-expert": [114], "collaboration": [115], "via": [116], "an": [117, 133, 140], "explicit": [118], "sub-query": [119, 138, 231], "dependency": [120, 149], "graph.": [121], "Comp-LLM": [122, 190, 223], "consists": [123], "three": [125], "components:": [126], "(1)": [127], "A": [128, 152, 174], "Sub-query": [129], "Generator": [130], "decomposes": [132], "input": [134], "query,": [135], "assigns": [136], "each": [137], "appropriate": [141], "expert": [142, 180], "using": [143], "embedding": [144], "similarity,": [145], "constructs": [147], "graph;": [150], "(2)": [151], "Query": [153], "Executor": [154], "processes": [156], "nodes": [157], "graph": [160], "identifies": [162], "opportunities": [163], "for": [164], "parallelism": [165], "based": [166], "dependencies": [168], "resource": [170], "constraints;": [171], "(3)": [173], "Response": [175], "Aggregator": [176], "synthesizes": [178], "intermediate": [179], "responses": [181], "coherent": [184], "final": [185], "answer.": [186], "Across": [187], "several": [188], "benchmarks,": [189], "achieves": [191], "up": [192], "11.01%": [194], "accuracy": [195], "improvement": [196, 227], "over": [197], "monolithic": [198], "LLMs": [199], "similar": [201], "size,": [202], "while": [203], "offering": [204], "1.67x--3.56x": [205], "reduction": [206], "size": [209], "no": [211], "degradation": [213], "relative": [214], "largest": [217], "its": [220], "family.": [221], "Additionally,": [222], "provides": [224], "1.1x--1.7x": [225], "latency": [226], "compared": [228], "processing.": [232]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.701200008392334, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.701200008392334, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.2451000064611435, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.013399999588727951, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2511.22955", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2511.22955", "pdf_url": "https://arxiv.org/pdf/2511.22955", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4408771256", "doi": "https://doi.org/10.2139/ssrn.5107732", "title": "The Mathematics of DeepSeek-R1: Theoretical Foundations and Comparative Analysis", "publication_year": 2025, "publication_date": "2025-01-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5007301330", "display_name": "Miquel Noguer I Alonso", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Miquel Noguer I Alonso", "raw_affiliation_strings": ["Artificial Intelligence in Finance Institute", "Artificial Intelligence in Finance Institute, New York, United States"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Artificial Intelligence in Finance Institute", "institution_ids": []}, {"raw_affiliation_string": "Artificial Intelligence in Finance Institute, New York, United States", "institution_ids": []}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5181000232696533, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5181000232696533, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.11739999800920486, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10462", "display_name": "Reinforcement Learning in Robotics", "score": 0.10980000346899033, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2139/ssrn.5107732", "is_oa": true, "landing_page_url": "https://doi.org/10.2139/ssrn.5107732", "pdf_url": null, "source": {"id": "https://openalex.org/S4210172589", "display_name": "SSRN Electronic Journal", "issn_l": "1556-5068", "issn": ["1556-5068"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I1318003438", "host_organization_name": "RELX Group (Netherlands)", "host_organization_lineage": ["https://openalex.org/I1318003438"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415256024", "doi": "https://doi.org/10.2139/ssrn.5584450", "title": "Beyond WEIRD Bias: Why Token-Based Legal AI Cannot Model Post-Colonial Compliance Dynamics", "publication_year": 2025, "publication_date": "2025-01-01", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5138243314", "display_name": "Ignacio Adri\u00e1n LERER", "orcid": "https://orcid.org/0009-0007-6378-9749", "observed_orcids": ["https://orcid.org/0009-0007-6378-9749"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Ignacio Adrian Lerer", "raw_affiliation_strings": ["Independent Researcher"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Independent Researcher", "institution_ids": []}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T13643", "display_name": "Artificial Intelligence in Law", "score": 0.9653000235557556, "subfield": {"id": "https://openalex.org/subfields/3320", "display_name": "Political Science and International Relations"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T13643", "display_name": "Artificial Intelligence in Law", "score": 0.9653000235557556, "subfield": {"id": "https://openalex.org/subfields/3320", "display_name": "Political Science and International Relations"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T10883", "display_name": "Ethics and Social Impacts of AI", "score": 0.015599999576807022, "subfield": {"id": "https://openalex.org/subfields/3311", "display_name": "Safety Research"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T11762", "display_name": "Law, Economics, and Judicial Systems", "score": 0.0019000000320374966, "subfield": {"id": "https://openalex.org/subfields/2002", "display_name": "Economics and Econometrics"}, "field": {"id": "https://openalex.org/fields/20", "display_name": "Economics, Econometrics and Finance"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "doi:10.2139/ssrn.5584450", "is_oa": true, "landing_page_url": "https://doi.org/10.2139/ssrn.5584450", "pdf_url": null, "source": {"id": "https://openalex.org/S4210172589", "display_name": "SSRN Electronic Journal", "issn_l": "1556-5068", "issn": ["1556-5068"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I1318003438", "host_organization_name": "RELX Group (Netherlands)", "host_organization_lineage": ["https://openalex.org/I1318003438"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4407424218", "doi": "https://doi.org/10.48550/arxiv.2502.07222", "title": "A Memory Efficient Randomized Subspace Optimization Method for Training Large Language Models", "publication_year": 2025, "publication_date": "2025-02-11", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100333361", "display_name": "Yiming Chen", "orcid": "https://orcid.org/0000-0002-8894-2902", "observed_orcids": ["https://orcid.org/0000-0002-8894-2902", "https://orcid.org/0000-0002-1861-3230", "https://orcid.org/0009-0001-8127-3670"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chen, Yiming", "raw_affiliation_strings": ["Beijing International Center for Mathematical Research , Peking University , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing International Center for Mathematical Research , Peking University , Beijing , China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100368731", "display_name": "Yuan Zhang", "orcid": "https://orcid.org/0000-0002-7849-208X", "observed_orcids": ["https://orcid.org/0000-0002-7849-208X"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhang, Yuan", "raw_affiliation_strings": ["Center for Data Science , Peking University , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Center for Data Science , Peking University , Beijing , China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100540864", "display_name": "Yin Liu", "orcid": "https://orcid.org/0009-0004-4044-958X", "observed_orcids": ["https://orcid.org/0009-0004-4044-958X"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Liu, Yin", "raw_affiliation_strings": ["Beijing International Center for Mathematical Research , Peking University , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing International Center for Mathematical Research , Peking University , Beijing , China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Yuan, Kun", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yuan, Kun", "raw_affiliation_strings": ["Center for Machine Learning Research , Peking University , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Center for Machine Learning Research , Peking University , Beijing , China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5006127137", "display_name": "Zaiwen Wen", "orcid": "https://orcid.org/0000-0003-1762-0671", "observed_orcids": ["https://orcid.org/0000-0003-1762-0671"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wen, Zaiwen", "raw_affiliation_strings": ["Beijing International Center for Mathematical Research , Peking University , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing International Center for Mathematical Research , Peking University , Beijing , China", "institution_ids": ["https://openalex.org/I20231570"]}]}], "abstract_inverted_index": {"The": [0], "memory": [1, 43, 55, 138, 170], "challenges": [2], "associated": [3], "with": [4, 31], "training": [5, 103], "Large": [6], "Language": [7], "Models": [8], "(LLMs)": [9], "have": [10, 28], "become": [11], "a": [12, 36, 88, 106, 114], "critical": [13], "concern,": [14], "particularly": [15], "when": [16], "using": [17], "the": [18, 42, 54, 81, 101, 120, 163, 168], "Adam": [19], "optimizer.": [20], "To": [21], "address": [22], "this": [23, 84], "issue,": [24], "numerous": [25], "memory-efficient": [26], "techniques": [27], "been": [29], "proposed,": [30], "GaLore": [32, 181], "standing": [33], "out": [34], "as": [35], "notable": [37], "example": [38], "designed": [39], "to": [40, 135, 161, 180], "reduce": [41, 137], "footprint": [44], "of": [45, 108, 174], "optimizer": [46, 144], "states.": [47, 145], "However,": [48], "these": [49], "approaches": [50], "do": [51], "not": [52, 78], "alleviate": [53], "burden": [56], "imposed": [57], "by": [58], "activations,": [59], "rendering": [60], "them": [61], "unsuitable": [62], "for": [63, 93, 140, 154], "scenarios": [64], "involving": [65], "long": [66], "context": [67], "sequences": [68], "or": [69], "large": [70], "mini-batches.": [71], "Moreover,": [72], "their": [73], "convergence": [74, 149], "properties": [75], "are": [76, 125], "still": [77], "well-understood": [79], "in": [80, 130], "literature.": [82], "In": [83], "work,": [85], "we": [86], "introduce": [87], "Randomized": [89], "Subspace": [90], "Optimization": [91], "framework": [92], "pre-training": [94], "and": [95, 119, 143, 151, 171, 182], "fine-tuning": [96], "LLMs.": [97], "Our": [98], "approach": [99], "decomposes": [100], "high-dimensional": [102], "problem": [104], "into": [105], "series": [107], "lower-dimensional": [109], "subproblems.": [110, 164], "At": [111], "each": [112], "iteration,": [113], "random": [115], "subspace": [116, 124], "is": [117], "selected,": [118], "parameters": [121], "within": [122], "that": [123], "optimized.": [126], "This": [127], "structured": [128], "reduction": [129], "dimensionality": [131], "allows": [132], "our": [133, 175], "method": [134], "simultaneously": [136], "usage": [139], "both": [141], "activations": [142], "We": [146], "establish": [147], "comprehensive": [148], "guarantees": [150], "derive": [152], "rates": [153], "various": [155], "scenarios,": [156], "accommodating": [157], "different": [158], "optimization": [159], "strategies": [160], "solve": [162], "Extensive": [165], "experiments": [166], "validate": [167], "superior": [169], "communication": [172], "efficiency": [173], "method,": [176], "achieving": [177], "performance": [178], "comparable": [179], "Adam.": [183]}, "primary_topic": {"id": "https://openalex.org/T11612", "display_name": "Stochastic Gradient Optimization Techniques", "score": 0.9672999978065491, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11612", "display_name": "Stochastic Gradient Optimization Techniques", "score": 0.9672999978065491, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.021299999207258224, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0012000000569969416, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.07222", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.07222", "pdf_url": "https://arxiv.org/pdf/2502.07222", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417250113", "doi": "https://doi.org/10.48550/arxiv.2504.07491", "title": "Kimi-VL Technical Report", "publication_year": 2025, "publication_date": "2025-04-10", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5116011620", "display_name": "Kimi Team", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kimi Team", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5088792207", "display_name": "Angang Du", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Du, Angang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Yin, Bohong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yin, Bohong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5028438544", "display_name": "Bowei Xing", "orcid": "https://orcid.org/0009-0001-3254-3902", "observed_orcids": ["https://orcid.org/0009-0001-3254-3902"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xing, Bowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Qu, Bowen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qu, Bowen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100412535", "display_name": "Bowen Wang", "orcid": "https://orcid.org/0000-0001-5794-2144", "observed_orcids": ["https://orcid.org/0000-0001-5794-2144"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Bowen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101559656", "display_name": "Cheng Chen", "orcid": "https://orcid.org/0000-0002-0586-0646", "observed_orcids": ["https://orcid.org/0000-0002-0586-0646"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Cheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103265154", "display_name": "Chenlin Zhang", "orcid": "https://orcid.org/0000-0001-5116-4046", "observed_orcids": ["https://orcid.org/0000-0001-5116-4046"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Chenlin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5044355459", "display_name": "Chenzhuang Du", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Du, Chenzhuang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wei, Chu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wei, Chu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111099880", "display_name": "Congcong Wang", "orcid": "https://orcid.org/0009-0006-5330-4691", "observed_orcids": ["https://orcid.org/0009-0006-5330-4691"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Congcong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5046162247", "display_name": "Dehao Zhang", "orcid": "https://orcid.org/0009-0007-4493-5595", "observed_orcids": ["https://orcid.org/0009-0007-4493-5595"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Dehao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Du, Dikang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Du, Dikang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100689748", "display_name": "Dongliang Wang", "orcid": "https://orcid.org/0000-0001-7052-4182", "observed_orcids": ["https://orcid.org/0000-0001-7052-4182"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Dongliang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5073001520", "display_name": "Enming Yuan", "orcid": "https://orcid.org/0000-0003-0755-9109", "observed_orcids": ["https://orcid.org/0000-0003-0755-9109"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yuan, Enming", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049660040", "display_name": "Enzhe Lu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lu, Enzhe", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020047491", "display_name": "Fang Li", "orcid": "https://orcid.org/0000-0002-6908-7920", "observed_orcids": ["https://orcid.org/0000-0002-6908-7920"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Fang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5065232789", "display_name": "Flood Sung", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sung, Flood", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wei, Guangda", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wei, Guangda", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015579277", "display_name": "Guokun Lai", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lai, Guokun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063667378", "display_name": "Zhu Han", "orcid": "https://orcid.org/0000-0002-6606-5822", "observed_orcids": ["https://orcid.org/0000-0002-6606-5822"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Han", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101528531", "display_name": "Hao Ding", "orcid": "https://orcid.org/0000-0003-2688-3852", "observed_orcids": ["https://orcid.org/0000-0003-2688-3852"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ding, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5046973045", "display_name": "Hao Hu", "orcid": "https://orcid.org/0000-0002-3567-609X", "observed_orcids": ["https://orcid.org/0000-0002-3567-609X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hu, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101457978", "display_name": "Hao Yang", "orcid": "https://orcid.org/0009-0005-8623-8531", "observed_orcids": ["https://orcid.org/0009-0005-8623-8531"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5082017754", "display_name": "Hao Zhang", "orcid": "https://orcid.org/0000-0002-0980-8888", "observed_orcids": ["https://orcid.org/0000-0002-0980-8888"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5039135087", "display_name": "Haoning Wu", "orcid": "https://orcid.org/0009-0001-8717-338X", "observed_orcids": ["https://orcid.org/0009-0001-8717-338X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Haoning", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5071838616", "display_name": "Haotian Yao", "orcid": "https://orcid.org/0000-0003-4674-1552", "observed_orcids": ["https://orcid.org/0000-0003-4674-1552"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yao, Haotian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083840550", "display_name": "H. Q. Lu", "orcid": "https://orcid.org/0000-0001-7759-9443", "observed_orcids": ["https://orcid.org/0000-0001-7759-9443"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lu, Haoyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100453945", "display_name": "Heng Wang", "orcid": "https://orcid.org/0000-0001-9338-6624", "observed_orcids": ["https://orcid.org/0000-0001-9338-6624"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Heng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110370692", "display_name": "Hongcheng Gao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gao, Hongcheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111649512", "display_name": "Huabin Zheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zheng, Huabin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100326742", "display_name": "Jiaming Li", "orcid": "https://orcid.org/0000-0002-0109-0867", "observed_orcids": ["https://orcid.org/0000-0002-0109-0867", "https://orcid.org/0000-0002-6424-1554"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Jiaming", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5078222384", "display_name": "Jianlin Su", "orcid": "https://orcid.org/0000-0002-1000-8458", "observed_orcids": ["https://orcid.org/0000-0002-1000-8458"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Su, Jianlin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5044720245", "display_name": "Jianzhou Wang", "orcid": "https://orcid.org/0000-0001-9078-7617", "observed_orcids": ["https://orcid.org/0000-0001-9078-7617", "https://orcid.org/0000-0001-6070-9888"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Jianzhou", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5062194013", "display_name": "Jiaqi Deng", "orcid": "https://orcid.org/0009-0001-1741-9882", "observed_orcids": ["https://orcid.org/0009-0001-1741-9882"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Deng, Jiaqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083333630", "display_name": "Jiezhong Qiu", "orcid": "https://orcid.org/0000-0001-9514-0708", "observed_orcids": ["https://orcid.org/0000-0001-9514-0708"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qiu, Jiezhong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5035415865", "display_name": "Jin Xie", "orcid": "https://orcid.org/0000-0001-8523-6052", "observed_orcids": ["https://orcid.org/0000-0001-8523-6052"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xie, Jin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100775334", "display_name": "Jinhong Wang", "orcid": "https://orcid.org/0000-0002-0809-9856", "observed_orcids": ["https://orcid.org/0000-0002-0809-9856"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Jinhong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Liu, Jingyuan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Jingyuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103205656", "display_name": "Junjie Yan", "orcid": "https://orcid.org/0009-0001-7918-3801", "observed_orcids": ["https://orcid.org/0009-0001-7918-3801"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yan, Junjie", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Ouyang, Kun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ouyang, Kun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100604810", "display_name": "Liang Chen", "orcid": "https://orcid.org/0000-0001-5006-4094", "observed_orcids": ["https://orcid.org/0000-0001-5006-4094"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Liang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Sui, Lin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sui, Lin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049419584", "display_name": "Longhui Yu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Longhui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5072164184", "display_name": "Mengfan Dong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Mengfan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Dong, Mengnan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Mengnan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101486827", "display_name": "Nuo Xu", "orcid": "https://orcid.org/0000-0003-2586-5804", "observed_orcids": ["https://orcid.org/0000-0003-2586-5804"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Nuo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110618868", "display_name": "Pengyu Cheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cheng, Pengyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Gu, Qizheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gu, Qizheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113209710", "display_name": "Runjie Zhou", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Runjie", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102884727", "display_name": "Shaowei Liu", "orcid": "https://orcid.org/0009-0004-6105-7999", "observed_orcids": ["https://orcid.org/0009-0004-6105-7999"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Shaowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Cao, Sihan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cao, Sihan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100596110", "display_name": "Tao Yu", "orcid": "https://orcid.org/0009-0001-8083-2418", "observed_orcids": ["https://orcid.org/0009-0001-8083-2418"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Tao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5091391590", "display_name": "Tianhui Song", "orcid": "https://orcid.org/0009-0007-4256-6407", "observed_orcids": ["https://orcid.org/0009-0007-4256-6407"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Song, Tianhui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Bai, Tongtong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bai, Tongtong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5019370942", "display_name": "Wei Song", "orcid": "https://orcid.org/0000-0002-4780-0350", "observed_orcids": ["https://orcid.org/0000-0002-4780-0350"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Song, Wei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "He, Weiran", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "He, Weiran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5013900180", "display_name": "Weixiao Huang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Weixiao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109953055", "display_name": "Weixin Xu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, Weixin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5071207563", "display_name": "Xiaokun Yuan", "orcid": "https://orcid.org/0000-0003-1515-8946", "observed_orcids": ["https://orcid.org/0000-0003-1515-8946"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yuan, Xiaokun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5065739587", "display_name": "Xingcheng Yao", "orcid": "https://orcid.org/0000-0001-5065-3016", "observed_orcids": ["https://orcid.org/0000-0001-5065-3016"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yao, Xingcheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5086375722", "display_name": "Xingzhe Wu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Xingzhe", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5038731476", "display_name": "Xinhao Li", "orcid": "https://orcid.org/0000-0002-6819-5460", "observed_orcids": ["https://orcid.org/0000-0002-6819-5460"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Xinhao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5028097242", "display_name": "Xinxing Zu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zu, Xinxing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5074959378", "display_name": "Xinyu Zhou", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Xinyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5054558475", "display_name": "Xinyuan Wang", "orcid": "https://orcid.org/0000-0002-8572-0407", "observed_orcids": ["https://orcid.org/0000-0002-8572-0407"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Xinyuan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5076124480", "display_name": "Yann Philippe Charles", "orcid": "https://orcid.org/0000-0002-0606-211X", "observed_orcids": ["https://orcid.org/0000-0002-0606-211X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Charles, Y.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110992737", "display_name": "Yan Zhong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhong, Yan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100460352", "display_name": "Li Yang", "orcid": "https://orcid.org/0000-0003-4791-1349", "observed_orcids": ["https://orcid.org/0000-0003-4791-1349"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Yang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101075955", "display_name": "Yangyang Hu", "orcid": "https://orcid.org/0009-0007-4457-5630", "observed_orcids": ["https://orcid.org/0009-0007-4457-5630"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Hu, Yangyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100695892", "display_name": "Yanru Chen", "orcid": "https://orcid.org/0000-0002-6531-7995", "observed_orcids": ["https://orcid.org/0000-0002-6531-7995", "https://orcid.org/0009-0003-4135-0488", "https://orcid.org/0009-0004-9330-1028", "https://orcid.org/0009-0007-1647-8999"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Yanru", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5104243967", "display_name": "Yejie Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yejie", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Liu, Yibo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Yibo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045257498", "display_name": "Yibo Miao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Miao, Yibo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Qin, Yidao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Qin, Yidao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100377728", "display_name": "Yimin Chen", "orcid": "https://orcid.org/0000-0003-1562-1443", "observed_orcids": ["https://orcid.org/0000-0003-1562-1443", "https://orcid.org/0009-0003-8204-7143"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Yimin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5143926194", "display_name": "Yiping Bao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bao, Yiping", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100662108", "display_name": "Yiqin Wang", "orcid": "https://orcid.org/0000-0002-2783-8861", "observed_orcids": ["https://orcid.org/0000-0002-2783-8861", "https://orcid.org/0000-0001-8480-2040"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yiqin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100931742", "display_name": "Yongsheng Kang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kang, Yongsheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5044541656", "display_name": "Yuanxin Liu", "orcid": "https://orcid.org/0000-0002-0673-4887", "observed_orcids": ["https://orcid.org/0000-0002-0673-4887"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Yuanxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101315785", "display_name": "Yuhao Dong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Yuhao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5012998749", "display_name": "Yulun Du", "orcid": "https://orcid.org/0009-0002-6171-6750", "observed_orcids": ["https://orcid.org/0009-0002-6171-6750"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Du, Yulun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wu, Yuxin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Yuxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100669046", "display_name": "Yuzhi Wang", "orcid": "https://orcid.org/0000-0002-9339-2386", "observed_orcids": ["https://orcid.org/0000-0002-9339-2386"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yuzhi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Yan, Yuzi", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yan, Yuzi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113154912", "display_name": "Zaida Zhou", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Zaida", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5008593161", "display_name": "Zhaowei Li", "orcid": "https://orcid.org/0000-0002-5176-4655", "observed_orcids": ["https://orcid.org/0000-0002-5176-4655"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Zhaowei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5012056827", "display_name": "Zhejun Jiang", "orcid": "https://orcid.org/0009-0000-0891-7380", "observed_orcids": ["https://orcid.org/0009-0000-0891-7380"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jiang, Zhejun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100459108", "display_name": "Zheng Zhang", "orcid": "https://orcid.org/0000-0002-2338-6108", "observed_orcids": ["https://orcid.org/0000-0002-2338-6108"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Zheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5050303775", "display_name": "Zhilin Yang", "orcid": "https://orcid.org/0000-0003-0944-4903", "observed_orcids": ["https://orcid.org/0000-0003-0944-4903", "https://orcid.org/0000-0002-8598-922X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Zhilin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103191012", "display_name": "Zhiqi Huang", "orcid": "https://orcid.org/0000-0002-8317-0966", "observed_orcids": ["https://orcid.org/0000-0002-8317-0966"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Zhiqi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Huang, Zihao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Zihao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102012495", "display_name": "Zijia Zhao", "orcid": "https://orcid.org/0009-0005-8242-6092", "observed_orcids": ["https://orcid.org/0009-0005-8242-6092"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhao, Zijia", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049260052", "display_name": "Ziwei Chen", "orcid": "https://orcid.org/0000-0001-7679-9582", "observed_orcids": ["https://orcid.org/0000-0001-7679-9582"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Ziwei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Lin, Zongyu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Lin, Zongyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0], "present": [1], "Kimi-VL,": [2, 168], "an": [3, 171], "efficient": [4, 87], "open-source": [5], "Mixture-of-Experts": [6], "(MoE)": [7], "vision-language": [8], "model": [9, 190], "(VLM)": [10], "that": [11], "offers": [12], "advanced": [13, 172], "multimodal": [14], "reasoning,": [15, 75], "long-context": [16], "understanding,": [17], "and": [18, 70, 76, 93, 109, 131, 145, 154, 184, 217], "strong": [19, 36, 192], "agent": [20, 49], "capabilities": [21, 60, 195], "-": [22], "all": [23], "while": [24, 95, 158, 211], "activating": [25], "only": [26], "2.8B": [27], "parameters": [28], "in": [29, 47, 98, 105], "its": [30], "language": [31, 65], "decoder": [32], "(Kimi-VL-A3B).": [33], "Kimi-VL": [34, 45, 102, 118], "demonstrates": [35], "performance": [37], "across": [38, 61], "challenging": [39, 63], "domains:": [40], "as": [41, 90], "a": [42, 113], "general-purpose": [43], "VLM,": [44], "excels": [46], "multi-turn": [48], "tasks": [50], "(e.g.,": [51], "OSWorld),": [52], "matching": [53], "flagship": [54], "models.": [55], "Furthermore,": [56], "it": [57, 82, 142], "exhibits": [58, 191], "remarkable": [59], "diverse": [62, 121], "vision": [64, 137], "tasks,": [66], "including": [67], "college-level": [68], "image": [69], "video": [71], "comprehension,": [72], "OCR,": [73], "mathematical": [74], "multi-image": [77], "understanding.": [78], "In": [79], "comparative": [80], "evaluations,": [81], "effectively": [83], "competes": [84], "with": [85], "cutting-edge": [86], "VLMs": [88], "such": [89], "GPT-4o-mini,": [91], "Qwen2.5-VL-7B,": [92], "Gemma-3-12B-IT,": [94], "surpassing": [96], "GPT-4o": [97], "several": [99], "key": [100], "domains.": [101], "also": [103], "advances": [104], "processing": [106], "long": [107, 122, 178], "contexts": [108], "perceiving": [110], "clearly.": [111], "With": [112], "128K": [114], "extended": [115], "context": [116], "window,": [117], "can": [119], "process": [120], "inputs,": [123, 149], "achieving": [124, 150], "impressive": [125], "scores": [126], "of": [127], "64.5": [128], "on": [129, 133, 152, 156, 197, 200, 203, 206, 209], "LongVideoBench": [130], "35.1": [132], "MMLongBench-Doc.": [134], "Its": [135], "native-resolution": [136], "encoder,": [138], "MoonViT,": [139], "further": [140], "allows": [141], "to": [143], "see": [144], "understand": [146], "ultra-high-resolution": [147], "visual": [148], "83.2": [151], "InfoVQA": [153], "34.5": [155], "ScreenSpot-Pro,": [157], "maintaining": [159], "lower": [160], "computational": [161], "cost": [162], "for": [163], "common": [164], "tasks.": [165], "Building": [166], "upon": [167], "we": [169], "introduce": [170], "long-thinking": [173], "variant:": [174], "Kimi-VL-Thinking-2506.": [175], "Developed": [176], "through": [177], "chain-of-thought": [179], "(CoT)": [180], "supervised": [181], "fine-tuning": [182], "(SFT)": [183], "reinforcement": [185], "learning": [186], "(RL),": [187], "the": [188], "latest": [189], "long-horizon": [193], "reasoning": [194], "(64.0": [196], "MMMU,": [198], "46.3": [199], "MMMU-Pro,": [201], "56.9": [202], "MathVision,": [204], "80.1": [205], "MathVista,": [207], "65.2": [208], "VideoMMMU)": [210], "obtaining": [212], "robust": [213], "general": [214], "abilities.": [215], "Code": [216], "models": [218], "are": [219], "publicly": [220], "accessible": [221], "at": [222], "https://github.com/MoonshotAI/Kimi-VL.": [223]}, "primary_topic": {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2504.07491", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2504.07491", "pdf_url": "https://arxiv.org/pdf/2504.07491", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 3, "is_retracted": false}, {"id": "https://openalex.org/W7124429279", "doi": "https://doi.org/10.48550/arxiv.2601.09714", "title": "Evaluating Novelty in AI-Generated Research Plans Using Multi-Workflow LLM Pipelines", "publication_year": 2025, "publication_date": "2025-12-24", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5123195713", "display_name": "Devesh Saraogi", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Saraogi, Devesh", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5062950793", "display_name": "Rohit Singhee", "orcid": "https://orcid.org/0000-0002-4891-7809", "observed_orcids": ["https://orcid.org/0000-0002-4891-7809"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Singhee, Rohit", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101298770", "display_name": "Dhruv Kumar", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kumar, Dhruv", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0], "integration": [1], "of": [2, 19, 119], "Large": [3], "Language": [4], "Models": [5], "(LLMs)": [6], "into": [7], "the": [8, 15, 146], "scientific": [9], "ecosystem": [10], "raises": [11], "fundamental": [12], "questions": [13], "about": [14], "creativity": [16], "and": [17, 57, 65, 92, 107, 113], "originality": [18], "AI-generated": [20], "research.": [21], "Recent": [22], "work": [23], "has": [24], "identified": [25], "``smart": [26], "plagiarism''": [27], "as": [28], "a": [29], "concern": [30], "in": [31], "single-step": [32], "prompting": [33], "approaches,": [34], "where": [35], "models": [36], "reproduce": [37], "existing": [38], "ideas": [39], "with": [40, 135], "terminological": [41], "shifts.": [42], "This": [43], "paper": [44], "investigates": [45], "whether": [46], "agentic": [47, 152], "workflows": [48, 115, 137, 153], "--": [49, 60], "multi-step": [50], "systems": [51], "employing": [52], "iterative": [53, 75], "reasoning,": [54], "evolutionary": [55, 80], "search,": [56], "recursive": [58, 90], "decomposition": [59], "can": [61, 154], "generate": [62], "more": [63], "novel": [64], "feasible": [66], "research": [67, 133, 157], "plans.": [68], "We": [69], "benchmark": [70], "five": [71], "reasoning": [72], "architectures:": [73], "Reflection-based": [74], "refinement,": [76], "Sakana": [77], "AI": [78], "v2": [79], "algorithms,": [81], "Google": [82], "Co-Scientist": [83], "multi-agent": [84], "framework,": [85], "GPT": [86], "Deep": [87], "Research": [88], "(GPT-5.1)": [89], "decomposition,": [91], "Gemini~3": [93], "Pro": [94], "multimodal": [95], "long-context": [96, 114], "pipeline.": [97], "Using": [98], "evaluations": [99], "from": [100], "thirty": [101], "proposals": [102], "each": [103], "on": [104], "novelty,": [105], "feasibility,": [106], "impact,": [108], "we": [109], "find": [110], "that": [111, 148], "decomposition-based": [112], "achieve": [116], "mean": [117], "novelty": [118], "4.17/5,": [120], "while": [121], "reflection-based": [122], "approaches": [123], "score": [124], "significantly": [125], "lower": [126], "(2.33/5).": [127], "Results": [128], "reveal": [129], "varied": [130], "performance": [131], "across": [132], "domains,": [134], "high-performing": [136], "maintaining": [138], "feasibility": [139], "without": [140], "sacrificing": [141], "creativity.": [142], "These": [143], "findings": [144], "support": [145], "view": [147], "carefully": [149], "designed": [150], "multi-stage": [151], "advance": [155], "AI-assisted": [156], "ideation.": [158]}, "primary_topic": {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.426800012588501, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.426800012588501, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.1574999988079071, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10906", "display_name": "AI-based Problem Solving and Planning", "score": 0.048900000751018524, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2601.09714", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2601.09714", "pdf_url": "https://arxiv.org/pdf/2601.09714", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4410768311", "doi": "https://doi.org/10.20944/preprints202505.2064.v1", "title": "A Comprehensive Review of Qwen and DeepSeek LLMs: Architecture, Performance and Applications", "publication_year": 2025, "publication_date": "2025-05-27", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5108368607", "display_name": "Satyadhar Joshi", "orcid": "https://orcid.org/0009-0002-6011-5080", "observed_orcids": ["https://orcid.org/0009-0002-6011-5080"]}, "institutions": [{"id": "https://openalex.org/I13955877", "display_name": "Bar-Ilan University", "ror": "https://ror.org/03kgsv495", "country_code": "IL", "type": "education", "lineage": ["https://openalex.org/I13955877"]}], "countries": ["IL"], "is_corresponding": true, "raw_author_name": "Satyadhar Joshi", "raw_affiliation_strings": ["Independent , Alumnus , International MBA , Bar-Ilan University , Israel;"], "raw_orcid": "https://orcid.org/0009-0002-6011-5080", "affiliations": [{"raw_affiliation_string": "Independent , Alumnus , International MBA , Bar-Ilan University , Israel;", "institution_ids": ["https://openalex.org/I13955877"]}]}], "abstract_inverted_index": {"This": [0, 32, 212], "study": [1], "analyzes": [2], "the": [3, 39, 57, 79, 192, 197, 239], "mid-2025": [4], "LLM": [5, 77, 194], "landscape": [6], "by": [7], "benchmarking": [8], "both": [9], "open-source": [10, 76, 100, 209], "and": [11, 21, 29, 41, 53, 67, 96, 101, 128, 139, 149, 164, 183, 205, 217, 222, 234], "proprietary": [12, 102], "models": [13], "across": [14, 112], "key": [15, 73], "dimensions": [16, 207], "including": [17, 157, 196], "performance,": [18], "computational": [19], "efficiency,": [20, 89], "operational": [22], "cost,": [23], "offering": [24], "insights": [25], "into": [26], "current": [27], "trade-offs": [28], "deployment": [30, 153], "strategies.": [31], "paper": [33, 187], "presents": [34], "a": [35, 225], "comprehensive": [36], "review": [37, 108], "of": [38, 64, 81, 200, 208, 214], "Qwen": [40], "DeepSeek": [42], "large": [43], "language": [44], "model": [45, 94, 110, 232], "families,": [46], "analyzing": [47], "their": [48, 235], "architectural": [49], "innovations,": [50], "performance": [51], "characteristics,": [52], "practical": [54], "applications": [55, 179], "within": [56], "broader": [58], "AI": [59, 210], "landscape.": [60], "Through": [61], "systematic": [62], "evaluation": [63], "recent": [65], "research": [66], "industry": [68, 170], "benchmarks,": [69], "we": [70], "identify": [71], "several": [72], "trends": [74], "shaping": [75], "development:": [78], "rise": [80], "hybrid": [82], "mixture-of-experts": [83], "architectures": [84], "that": [85], "dramatically": [86], "improve": [87], "inference": [88], "breakthrough": [90], "techniques": [91], "enabling": [92], "cost-effective": [93], "customization,": [95], "growing": [97, 198], "parity": [98], "between": [99], "systems": [103], "in": [104, 136, 146, 180, 237], "specialized": [105, 201], "domains.": [106], "The": [107, 186], "compares": [109], "designs": [111], "multiple": [113], "scales,": [114], "from": [115], "7B": [116], "to": [117, 123], "235B": [118], "parameters,": [119], "highlighting": [120], "unique": [121], "approaches": [122], "memory": [124], "optimization,": [125], "context": [126], "handling,": [127], "multi-task": [129], "learning.": [130], "Performance": [131], "analysis": [132], "reveals": [133], "consistent": [134], "strengths": [135], "mathematical": [137], "reasoning": [138], "coding": [140], "tasks,": [141], "while": [142], "identifying": [143], "remaining": [144], "gaps": [145], "creative": [147], "generation": [148], "multilingual": [150], "capabilities.": [151], "Practical": [152], "considerations": [154], "are": [155, 172], "examined,": [156], "local": [158], "execution": [159], "optimizations,": [160], "document": [161], "processing": [162], "pipelines,": [163], "security": [165], "vulnerabilities.": [166], "Emerging": [167], "impacts": [168], "on": [169], "verticals": [171], "discussed,": [173], "with": [174, 189, 224], "case": [175], "studies": [176], "demonstrating": [177], "successful": [178], "finance,": [181], "healthcare,": [182], "scientific": [184], "research.": [185], "concludes": [188], "projections": [190], "about": [191], "evolving": [193], "ecosystem,": [195], "importance": [199], "models,": [202], "hardware-software": [203], "co-design,": [204], "geopolitical": [206], "development.": [211], "synthesis": [213], "technical": [215], "literature": [216], "empirical": [218], "results": [219], "provides": [220], "researchers": [221], "practitioners": [223], "unified": [226], "reference": [227], "for": [228], "understanding": [229], "these": [230], "influential": [231], "families": [233], "role": [236], "advancing": [238], "field.": [240]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.95169997215271, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.95169997215271, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.03830000013113022, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.002899999963119626, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.20944/preprints202505.2064.v1", "is_oa": true, "landing_page_url": "https://doi.org/10.20944/preprints202505.2064.v1", "pdf_url": "https://www.preprints.org/frontend/manuscript/fea14a68a2e533f3c500b66dcfc2d224/download_pub", "source": {"id": "https://openalex.org/S6309402219", "display_name": "Preprints.org", "issn_l": null, "issn": null, "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": null, "host_organization_name": null, "host_organization_lineage": ["https://openalex.org/P4310310987"], "host_organization_lineage_names": ["Multidisciplinary Digital Publishing Institute"], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 2, "is_retracted": false}, {"id": "https://openalex.org/W4413437971", "doi": "https://doi.org/10.62051/rs49my32", "title": "Research and Analysis on Text Interaction Methods Based on Large Language Models", "publication_year": 2025, "publication_date": "2025-07-10", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5103183194", "display_name": "Shyr-Shen Yu", "orcid": "https://orcid.org/0000-0001-5711-5018", "observed_orcids": ["https://orcid.org/0000-0001-5711-5018"]}, "institutions": [{"id": "https://openalex.org/I4210100789", "display_name": "Wuhan College", "ror": "https://ror.org/01dashf18", "country_code": "CN", "type": "nonprofit", "lineage": ["https://openalex.org/I4210100789"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Shangze Yu", "raw_affiliation_strings": ["Information Engineering College , Wuhan College , Wuhan , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Information Engineering College , Wuhan College , Wuhan , China", "institution_ids": ["https://openalex.org/I4210100789"]}]}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs),": [3], "as": [4, 89, 104, 159], "a": [5, 178, 197], "cutting-edge": [6], "technology": [7, 233], "in": [8, 45, 86, 131, 145], "the": [9, 46, 78, 119, 146, 194, 229], "field": [10], "of": [11, 36, 84, 148, 181, 196, 208, 232], "artificial": [12], "intelligence,": [13], "have": [14], "revolutionized": [15], "human-computer": [16], "interaction": [17, 70], "paradigms": [18], "by": [19, 122], "integrating": [20], "natural": [21], "language": [22], "processing,": [23], "Transformer": [24, 47], "architecture,": [25], "and": [26, 34, 39, 58, 67, 81, 95, 111, 143, 166, 189, 211, 227], "reinforcement": [27, 65], "learning": [28, 108], "techniques,": [29], "thereby": [30], "achieving": [31], "in-depth": [32], "understanding": [33], "generation": [35, 62], "linguistic": [37], "logic": [38], "emotions.": [40], "Their": [41], "technological": [42, 123, 184, 198], "core": [43], "lies": [44], "architecture": [48], "trained": [49], "on": [50, 206, 219], "massive": [51], "datasets,": [52], "which": [53], "can": [54], "capture": [55], "complex": [56], "semantics": [57], "contextual": [59], "associations,": [60], "optimize": [61], "quality": [63], "through": [64, 234], "learning,": [66], "enhance": [68], "interpersonal-like": [69], "via": [71], "affective": [72], "computing.": [73], "This": [74], "paper": [75], "systematically": [76], "reviews": [77], "technical": [79], "framework": [80, 200], "application": [82, 101], "scenarios": [83], "LLMs:": [85], "fields": [87], "such": [88, 103, 158], "intelligent": [90], "customer": [91], "service,": [92], "educational": [93], "assessment,": [94], "game": [96], "narration,": [97], "these": [98, 154], "models": [99], "demonstrate": [100], "values": [102], "multi-turn": [105], "dialogue,": [106], "personalized": [107], "path": [109], "planning,": [110], "dynamic": [112], "plot": [113], "generation.": [114], "Meanwhile,": [115], "it": [116], "delves": [117], "into": [118], "challenges": [120], "faced": [121], "development,": [124], "including": [125], "ethical": [126, 168], "risks": [127], "arising": [128], "from": [129, 183], "biases": [130], "training": [132], "data,": [133], "deployment": [134], "cost": [135], "issues": [136], "due": [137], "to": [138, 153, 186, 201], "high": [139], "computational": [140, 212], "power": [141, 213], "requirements,": [142], "deficiencies": [144], "controllability": [147], "generated": [149], "content.": [150], "In": [151], "response": [152], "issues,": [155], "collaborative": [156], "solutions": [157], "multimodal": [160], "data": [161], "fusion,": [162], "lightweight": [163], "model": [164, 221], "deployment,": [165], "interdisciplinary": [167], "governance": [169, 199], "are": [170, 176], "proposed.": [171], "Research": [172], "indicates": [173], "that": [174], "LLMs": [175], "at": [177], "critical": [179], "stage": [180], "transitioning": [182], "breakthroughs": [185], "large-scale": [187], "applications,": [188], "their": [190], "sustainable": [191], "development": [192, 231], "necessitates": [193], "construction": [195], "achieve": [202], "social": [203], "value": [204], "balance,": [205], "top": [207], "algorithm": [209], "optimization": [210], "enhancement.": [214], "Future": [215], "research": [216], "should": [217], "focus": [218], "enhancing": [220], "interpretability,": [222], "exploring": [223], "green": [224], "computing": [225], "pathways,": [226], "promoting": [228], "virtuous": [230], "human-machine": [235], "collaboration": [236], "mechanisms.": [237]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5702999830245972, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5702999830245972, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12031", "display_name": "Speech and dialogue systems", "score": 0.3361999988555908, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.05299999937415123, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.62051/rs49my32", "is_oa": true, "landing_page_url": "https://doi.org/10.62051/rs49my32", "pdf_url": "https://wepub.org/index.php/TCSISR/article/download/5454/5988", "source": {"id": "https://openalex.org/S4390297101", "display_name": "Transactions on Computer Science and Intelligent Systems Research", "issn_l": "2960-1800", "issn": ["2960-1800", "2960-2238"], "is_oa": true, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core"], "host_organization": null, "host_organization_name": null, "host_organization_lineage": [], "host_organization_lineage_names": [], "type": "journal"}, "license": "cc-by-nc", "license_id": "https://openalex.org/licenses/cc-by-nc", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Transactions on Computer Science and Intelligent Systems Research", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416432874", "doi": "https://doi.org/10.48550/arxiv.2511.00907", "title": "Transformers as Intrinsic Optimizers: Forward Inference through the Energy Principle", "publication_year": 2025, "publication_date": "2025-11-02", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5104187384", "display_name": "Ruifeng Ren", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ren, Ruifeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Ouyang, Sheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ouyang, Sheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5005691346", "display_name": "Huayi Tang", "orcid": "https://orcid.org/0009-0003-7864-8827", "observed_orcids": ["https://orcid.org/0009-0003-7864-8827"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Huayi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5028071674", "display_name": "Yong Liu", "orcid": "https://orcid.org/0000-0002-6739-621X", "observed_orcids": ["https://orcid.org/0000-0002-6739-621X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Yong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Attention-based": [0], "Transformers": [1], "have": [2, 13], "demonstrated": [3], "strong": [4], "adaptability": [5], "across": [6], "a": [7, 38, 55, 64, 157], "wide": [8], "range": [9], "of": [10, 17, 52, 71, 124, 170], "tasks": [11], "and": [12, 83, 101, 152], "become": [14], "the": [15, 50, 75, 79, 84, 136, 145, 168, 171], "backbone": [16], "modern": [18], "Large": [19], "Language": [20], "Models": [21], "(LLMs).": [22], "However,": [23], "their": [24, 110], "underlying": [25], "mechanisms": [26], "remain": [27], "open": [28], "for": [29, 41, 167, 174], "further": [30], "exploration.": [31], "The": [32], "energy-based": [33, 66, 122, 172], "perspective": [34], "has": [35], "long": [36], "provided": [37], "valuable": [39], "principle": [40, 51], "understanding": [42], "neural": [43], "computation.": [44], "In": [45], "this": [46, 114, 118], "paper,": [47], "we": [48, 120, 134], "revisit": [49], "energy": [53, 77, 81], "as": [54], "lens": [56], "to": [57, 144], "understand": [58], "attention-based": [59], "Transformer": [60], "models.": [61], "We": [62, 88], "present": [63], "unified": [65], "framework": [67, 173], "which": [68], "is": [69], "composed": [70], "three": [72], "key": [73], "components:": [74], "local": [76], "$E_i$,": [78], "global": [80], "$F$,": [82], "employed": [85], "optimization": [86], "algorithms.": [87], "show": [89], "that": [90], "different": [91], "attention": [92, 100, 104, 125, 138, 160, 176], "forms": [93], "including": [94], "unnormalized": [95], "linear": [96, 99], "attention,": [97], "gated": [98], "standard": [102, 142], "softmax": [103], "can": [105], "be": [106], "induced": [107], "by": [108, 128], "choosing": [109], "corresponding": [111, 158], "recipes": [112], "within": [113], "framework.": [115], "Building": [116], "on": [117, 141], "framework,": [119], "propose": [121], "modifications": [123], "structures.": [126], "Inspired": [127], "classical": [129], "gradient": [130], "descent": [131], "(GD)": [132], "algorithms,": [133], "extend": [135], "original": [137], "formulation": [139], "based": [140], "GD": [143], "momentum-based": [146], "GD,": [147], "Nesterov": [148], "Accelerated": [149], "Gradient": [150], "(NAG),": [151], "Newton's": [153], "method,": [154], "each": [155], "inducing": [156], "new": [159], "structure.": [161], "Our": [162], "experiments": [163], "provide": [164], "preliminary": [165], "support": [166], "potential": [169], "designing": [175], "mechanisms.": [177]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8633000254631042, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8633000254631042, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.02669999934732914, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10320", "display_name": "Neural Networks and Applications", "score": 0.017899999395012856, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2511.00907", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2511.00907", "pdf_url": "https://arxiv.org/pdf/2511.00907", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414581508", "doi": "https://doi.org/10.48550/arxiv.2505.18610", "title": "PM-KVQ: Progressive Mixed-precision KV Cache Quantization for Long-CoT LLMs", "publication_year": 2025, "publication_date": "2025-05-24", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5040810891", "display_name": "Tengxuan Liu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Tengxuan", "raw_affiliation_strings": ["Tsinghua University , 2 Infinigence-AI ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University , 2 Infinigence-AI ,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101827286", "display_name": "Shiyao Li", "orcid": "https://orcid.org/0000-0002-7258-0564", "observed_orcids": ["https://orcid.org/0000-0002-7258-0564"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Shiyao", "raw_affiliation_strings": ["Tsinghua University , 2 Infinigence-AI ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University , 2 Infinigence-AI ,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Yang, Jiayi", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I78577930", "display_name": "Columbia University", "ror": "https://ror.org/00hj8s172", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I78577930"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Yang, Jiayi", "raw_affiliation_strings": ["Columbia University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Columbia University ,", "institution_ids": ["https://openalex.org/I78577930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5039134666", "display_name": "Tianchen Zhao", "orcid": "https://orcid.org/0000-0002-2071-7514", "observed_orcids": ["https://orcid.org/0000-0002-2071-7514", "https://orcid.org/0000-0003-2911-7205", "https://orcid.org/0009-0003-6001-5669"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhao, Tianchen", "raw_affiliation_strings": ["Tsinghua University , 2 Infinigence-AI ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University , 2 Infinigence-AI ,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5064086456", "display_name": "Feng Zhou", "orcid": "https://orcid.org/0000-0002-2906-8127", "observed_orcids": ["https://orcid.org/0000-0002-2906-8127", "https://orcid.org/0009-0006-5301-7019"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhou, Feng", "raw_affiliation_strings": ["OPPO AI Center , Beijing , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "OPPO AI Center , Beijing , China,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Song, Xiaohui", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Song, Xiaohui", "raw_affiliation_strings": ["OPPO AI Center , Beijing , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "OPPO AI Center , Beijing , China,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102805465", "display_name": "Guohao Dai", "orcid": "https://orcid.org/0000-0002-8464-0130", "observed_orcids": ["https://orcid.org/0000-0002-8464-0130", "https://orcid.org/0000-0003-0849-3252"]}, "institutions": [{"id": "https://openalex.org/I183067930", "display_name": "Shanghai Jiao Tong University", "ror": "https://ror.org/0220qvk04", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I183067930"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Dai, Guohao", "raw_affiliation_strings": ["Shanghai Jiaotong University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Shanghai Jiaotong University", "institution_ids": ["https://openalex.org/I183067930"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5049910854", "display_name": "Shengen Yan", "orcid": "https://orcid.org/0009-0005-3858-7972", "observed_orcids": ["https://orcid.org/0009-0005-3858-7972"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yan, Shengen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103867707", "display_name": "Huazhong Yang", "orcid": "https://orcid.org/0000-0003-2421-353X", "observed_orcids": ["https://orcid.org/0000-0003-2421-353X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yang, Huazhong", "raw_affiliation_strings": ["Tsinghua University , 2 Infinigence-AI ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University , 2 Infinigence-AI ,", "institution_ids": []}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101879008", "display_name": "Yu Wang", "orcid": "https://orcid.org/0000-0003-0212-6830", "observed_orcids": ["https://orcid.org/0000-0003-0212-6830", "https://orcid.org/0000-0003-3511-0288"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Yu", "raw_affiliation_strings": ["Tsinghua University , 2 Infinigence-AI ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University , 2 Infinigence-AI ,", "institution_ids": []}]}], "abstract_inverted_index": {"Recently,": [0], "significant": [1, 64], "progress": [2], "has": [3, 40, 48], "been": [4, 49], "made": [5], "in": [6, 52, 128, 133, 152, 174], "developing": [7], "reasoning-capable": [8], "Large": [9, 74], "Language": [10], "Models": [11], "(LLMs)": [12], "through": [13], "long": [14], "Chain-of-Thought": [15], "(CoT)": [16], "techniques.": [17], "However,": [18, 55], "this": [19], "long-CoT": [20, 61, 145, 231], "reasoning": [21, 237], "process": [22], "imposes": [23], "substantial": [24], "memory": [25, 34, 181, 250], "overhead": [26], "due": [27, 67], "to": [28, 60, 68, 80, 97, 106, 119, 147, 166, 183, 188, 219, 242], "the": [29, 69, 89, 111, 122, 129, 149, 169, 196, 221, 248, 257], "large": [30, 98], "Key-Value": [31], "(KV)": [32], "Cache": [33, 38, 91, 141, 173], "overhead.": [35], "Post-training": [36], "KV": [37, 90, 140, 172], "quantization": [39, 100, 164], "emerged": [41], "as": [42], "a": [43, 162, 185, 204], "promising": [44], "compression": [45], "technique": [46], "and": [47, 85, 252], "extensively": [50], "studied": [51], "short-context": [53, 114], "scenarios.": [54], "directly": [56, 87], "applying": [57], "existing": [58], "methods": [59, 78], "LLMs": [62, 146, 232], "causes": [63], "performance": [65, 134, 239], "degradation": [66], "following": [70], "two": [71, 153], "reasons:": [72], "(1)": [73, 155], "cumulative": [75, 99, 158], "error:": [76], "Existing": [77], "fail": [79], "adequately": [81], "leverage": [82], "available": [83], "memory,": [84], "they": [86], "quantize": [88], "during": [92, 116], "each": [93, 175], "decoding": [94], "step,": [95], "leading": [96], "error.": [101], "(2)": [102, 193], "Short-context": [103], "calibration:": [104], "Due": [105], "Rotary": [107], "Positional": [108], "Embedding": [109], "(RoPE),": [110], "use": [112], "of": [113, 124, 171, 224], "data": [115, 215, 222], "calibration": [117, 197, 206, 214], "fails": [118], "account": [120], "for": [121, 144], "distribution": [123, 223], "less": [125], "frequent": [126], "channels": [127], "Key": [130], "Cache,": [131], "resulting": [132], "loss.": [135], "We": [136], "propose": [137, 179, 203], "Progressive": [138], "Mixed-Precision": [139], "Quantization": [142], "(PM-KVQ)": [143], "address": [148], "above": [150], "issues": [151], "folds:": [154], "To": [156, 194], "reduce": [157], "error,": [159], "we": [160, 178, 202], "design": [161], "progressive": [163], "strategy": [165, 207], "gradually": [167], "lower": [168], "bit-width": [170, 187], "block.": [176], "Then,": [177], "block-wise": [180], "allocation": [182], "assign": [184], "higher": [186], "more": [189], "sensitive": [190], "transformer": [191], "blocks.": [192], "increase": [195], "length": [198], "without": [199], "additional": [200], "overhead,": [201], "new": [205], "with": [208, 216], "positional": [209, 217], "interpolation": [210, 218], "that": [211, 234], "leverages": [212], "short": [213], "approximate": [220], "long-context": [225], "data.": [226], "Extensive": [227], "experiments": [228], "on": [229], "7B-70B": [230], "show": [233], "PM-KVQ": [235], "improves": [236], "benchmark": [238], "by": [240], "up": [241], "8%": [243], "over": [244, 256], "SOTA": [245], "baselines": [246], "under": [247], "same": [249], "budget": [251], "achieves": [253], "2.73-5.18x": [254], "throughput": [255], "original": [258], "16-bit": [259], "LLMs.": [260]}, "primary_topic": {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.8003000020980835, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.8003000020980835, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.05260000005364418, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.020999999716877937, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2505.18610", "is_oa": true, "landing_page_url": "https://arxiv.org/abs/2505.18610", "pdf_url": "https://arxiv.org/pdf/2505.18610", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417030020", "doi": "https://doi.org/10.48550/arxiv.2512.03494", "title": "A Preliminary Study on the Promises and Challenges of Native Top-$k$ Sparse Attention", "publication_year": 2025, "publication_date": "2025-12-03", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5123871991", "display_name": "Di Xiu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Xiu, Di", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109377270", "display_name": "Hongyin Tang", "orcid": "https://orcid.org/0009-0006-7745-7537", "observed_orcids": ["https://orcid.org/0009-0006-7745-7537"]}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Tang, Hongyin", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5043800885", "display_name": "Bolin Rong", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Rong, Bolin", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5120589688", "display_name": "Lizhi Yan", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yan, Lizhi", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5027638160", "display_name": "Jingang Wang", "orcid": "https://orcid.org/0000-0001-5758-0083", "observed_orcids": ["https://orcid.org/0000-0001-5758-0083"]}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Jingang", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5121800571", "display_name": "Yifan Lu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Lu, Yifan", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5068562468", "display_name": "Xunliang Cai", "orcid": "https://orcid.org/0000-0003-0685-9556", "observed_orcids": ["https://orcid.org/0000-0003-0685-9556"]}, "institutions": [{"id": "https://openalex.org/I4405259410", "display_name": "Meituan (China)", "ror": "https://ror.org/0272kpg92", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I4405259410"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Cai, Xunliang", "raw_affiliation_strings": ["Meituan , Beijing , China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Meituan , Beijing , China", "institution_ids": ["https://openalex.org/I4405259410"]}]}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs)": [3], "are": [4, 236], "increasingly": [5], "prevalent": [6], "in": [7, 226], "the": [8, 24, 41, 47, 53, 61, 75, 79, 83, 86, 90, 115, 125, 136, 150, 160, 189, 194, 205, 231], "field": [9], "of": [10, 26, 46, 63, 139, 154, 162, 188, 207, 223], "long-context": [11], "modeling,": [12], "however,": [13], "their": [14], "inference": [15, 130], "computational": [16, 152], "costs": [17], "have": [18], "become": [19], "a": [20, 37, 173, 201, 220], "critical": [21], "bottleneck": [22], "hindering": [23], "advancement": [25], "tasks": [27, 104], "such": [28, 105], "as": [29, 85, 106], "agents": [30], "and": [31, 43, 55, 108, 129, 180, 183], "multimodal": [32], "applications.": [33], "This": [34], "report": [35, 199], "conducts": [36], "preliminary": [38], "investigation": [39], "into": [40], "effectiveness": [42, 62], "theoretical": [44, 202], "mechanisms": [45], "Top-$k$": [48, 65, 117, 132, 140, 156, 164, 216, 240], "Attention": [49, 118, 133, 217], "mechanism": [50], "during": [51, 89], "both": [52], "decoding": [54, 91], "training": [56, 119, 128], "phases.": [57], "First,": [58], "we": [59, 112, 158, 184], "validate": [60], "exact": [64, 155], "Decoding": [66], "through": [67], "extensive": [68], "experimentation.": [69], "Experiments": [70, 121], "demonstrate": [71], "that": [72, 123, 212, 233], "retaining": [73], "only": [74], "pivotal": [76], "Keys": [77], "with": [78], "highest": [80], "similarity": [81], "to": [82, 215, 239], "Query": [84], "context": [87], "window": [88], "stage": [92], "achieves": [93], "performance": [94, 179], "comparable": [95], "to,": [96], "or": [97], "even": [98], "surpassing,": [99], "full": [100], "attention": [101], "on": [102, 167], "downstream": [103, 168, 177, 227], "HELMET": [107], "LongBench": [109], "v2.": [110], "Second,": [111], "further": [113, 137], "explore": [114], "native": [116], "strategy.": [120], "confirm": [122], "ensuring": [124], "consistency": [126], "between": [127, 176], "regarding": [131], "operations": [134], "facilitates": [135], "unlocking": [138], "Decoding's": [141], "potential,": [142], "thereby": [143], "significantly": [144], "enhancing": [145], "model": [146], "performance.": [147], "Furthermore,": [148], "considering": [149], "high": [151], "complexity": [153], "Attention,": [157], "investigate": [159], "impact": [161], "approximate": [163], "algorithm": [165], "precision": [166, 192], "tasks.": [169], "Our": [170], "research": [171], "confirms": [172], "positive": [174], "correlation": [175], "task": [178], "approximation": [181], "fidelity,": [182], "provide": [185], "statistical": [186], "evaluations": [187], "Lightning": [190], "Indexer's": [191], "within": [193], "DeepSeek-V3.2-Exp": [195], "model.": [196], "Finally,": [197], "this": [198], "provides": [200], "interpretation": [203], "from": [204], "perspective": [206], "Entropy.": [208], "Experimental": [209], "observations": [210], "indicate": [211], "models": [213], "subjected": [214], "SFT": [218], "exhibit": [219], "distinct": [221], "phenomenon": [222], "entropy": [224], "reduction": [225], "tasks,": [228], "which": [229], "validates": [230], "hypothesis": [232], "low-entropy": [234], "states": [235], "better": [237], "adapted": [238], "Decoding.": [241]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8057000041007996, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8057000041007996, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.17159999907016754, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.00419999985024333, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.03494", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.03494", "pdf_url": "https://arxiv.org/pdf/2512.03494", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415337488", "doi": "https://doi.org/10.48550/arxiv.2509.24899", "title": "Attention Surgery: An Efficient Recipe to Linearize Your Video Diffusion Transformer", "publication_year": 2025, "publication_date": "2025-09-29", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5083441393", "display_name": "Mohsen Ghafoorian", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ghafoorian, Mohsen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066772573", "display_name": "Denis Korzhenkov", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Korzhenkov, Denis", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5045620960", "display_name": "Amirhossein Habibian", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Habibian, Amirhossein", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Transformer-based": [0], "video": [1, 7], "diffusion": [2], "models": [3], "(VDMs)": [4], "deliver": [5], "state-of-the-art": [6, 122], "generation": [8], "quality": [9], "but": [10], "are": [11], "constrained": [12], "by": [13, 74], "the": [14, 39, 67], "quadratic": [15], "cost": [16], "of": [17, 41, 143], "self-attention,": [18], "making": [19], "long": [20], "sequences": [21], "and": [22, 89, 95, 113, 126, 131, 148], "high": [23], "resolutions": [24], "computationally": [25], "expensive.": [26], "While": [27], "linear": [28, 59, 90], "attention": [29, 43, 62, 86], "offers": [30], "sub-quadratic": [31], "complexity,": [32], "previous": [33], "approaches": [34], "have": [35], "failed": [36], "to": [37, 110, 118], "match": [38], "expressiveness": [40, 112], "softmax": [42, 88], "unless": [44], "retrained": [45], "at": [46], "significant": [47], "computational": [48], "cost.": [49], "We": [50], "introduce": [51], "Attention": [52, 136], "Surgery,": [53], "an": [54], "efficient": [55, 123], "framework": [56], "that": [57], "enables": [58], "or": [60], "hybrid": [61, 85], "in": [63, 77, 153], "pretrained": [64], "VDMs,": [65], "eliminating": [66], "need": [68], "for": [69, 156], "training": [70], "from": [71], "scratch.": [72], "Inspired": [73], "recent": [75], "advances": [76], "language": [78], "models,": [79], "our": [80], "method": [81], "combines": [82], "a": [83, 92, 100, 106, 121, 132], "novel": [84], "mechanism-mixing": [87], "tokens-with": [91], "lightweight": [93], "distillation": [94], "fine-tuning": [96], "pipeline": [97], "requiring": [98], "only": [99], "few": [101], "GPU-days.": [102], "Additionally,": [103], "we": [104], "incorporate": [105], "cost-aware": [107], "block-rate": [108], "strategy": [109], "balance": [111], "efficiency": [114], "across": [115], "layers.": [116], "Applied": [117], "Wan2.1": [119], "1.3B,": [120], "transformer": [124], "VDM": [125], "evaluated": [127], "on": [128], "VBench,": [129], "VBench2.0": [130], "human": [133], "preference": [134], "study,": [135], "Surgery": [137], "achieves": [138], "competitive": [139], "results.": [140], "Furthermore,": [141], "measurements": [142], "on-mobile": [144], "latency,": [145], "memory": [146], "usage,": [147], "FLOPs": [149], "demonstrate": [150], "notable": [151], "improvements": [152], "scaling": [154], "behavior": [155], "longer": [157], "videos.": [158], "Project": [159], "page": [160], "is": [161], "available": [162], "at:": [163], "https://qualcomm-ai-research.github.io/attention-surgery.": [164]}, "primary_topic": {"id": "https://openalex.org/T10775", "display_name": "Generative Adversarial Networks and Image Synthesis", "score": 0.9937000274658203, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10775", "display_name": "Generative Adversarial Networks and Image Synthesis", "score": 0.9937000274658203, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.003700000001117587, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.0008999999845400453, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2509.24899", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2509.24899", "pdf_url": "https://arxiv.org/pdf/2509.24899", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417068622", "doi": "https://doi.org/10.48550/arxiv.2503.23007", "title": "S2MoE: Robust Sparse Mixture of Experts via Stochastic Learning", "publication_year": 2025, "publication_date": "2025-03-29", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5035809896", "display_name": "Giang Do", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Do, Giang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Le, Hung", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Le, Hung", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5085471517", "display_name": "Truyen Tran", "orcid": "https://orcid.org/0000-0001-6531-8907", "observed_orcids": ["https://orcid.org/0000-0001-6531-8907"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tran, Truyen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Sparse": [0, 94], "Mixture": [1, 95], "of": [2, 8, 20, 31, 96, 106], "Experts": [3, 97], "(SMoE)": [4], "enables": [5], "efficient": [6], "training": [7, 23], "large": [9], "language": [10], "models": [11], "by": [12, 141], "routing": [13, 69, 134], "input": [14, 71], "tokens": [15], "to": [16, 28, 42, 64, 72, 79, 109, 132], "a": [17, 89, 104], "select": [18], "number": [19], "experts.": [21], "However,": [22], "SMoE": [24], "remains": [25], "challenging": [26], "due": [27], "the": [29, 40, 60, 73], "issue": [30], "representation": [32, 65], "collapse.": [33], "Recent": [34], "studies": [35], "have": [36], "focused": [37], "on": [38], "improving": [39], "router": [41], "mitigate": [43], "this": [44, 85], "problem,": [45], "but": [46], "existing": [47], "approaches": [48], "face": [49], "two": [50], "key": [51], "limitations:": [52], "(1)": [53], "expert": [54], "embeddings": [55], "are": [56], "significantly": [57], "smaller": [58], "than": [59], "model's": [61], "dimension,": [62], "contributing": [63], "collapse,": [66], "and": [67, 114], "(2)": [68], "each": [70], "Top-K": [74], "experts": [75, 107], "can": [76], "cause": [77], "them": [78], "learn": [80, 110], "overly": [81], "similar": [82], "features.": [83], "In": [84], "work,": [86], "we": [87], "propose": [88], "novel": [90], "approach": [91], "called": [92], "Robust": [93], "via": [98, 117], "Stochastic": [99], "Learning": [100, 118], "(S2MoE),": [101], "which": [102], "is": [103], "mixture": [105], "designed": [108], "from": [111], "both": [112], "deterministic": [113], "non-deterministic": [115], "inputs": [116], "under": [119], "Uncertainty.": [120], "Extensive": [121], "experiments": [122], "across": [123], "various": [124], "tasks": [125], "demonstrate": [126], "that": [127], "S2MoE": [128], "achieves": [129], "performance": [130], "comparable": [131], "other": [133], "methods": [135], "while": [136], "reducing": [137], "computational": [138], "inference": [139], "costs": [140], "28%.": [142]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5845000147819519, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5845000147819519, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.24789999425411224, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.10189999639987946, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.23007", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.23007", "pdf_url": "https://arxiv.org/pdf/2503.23007", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416042319", "doi": "https://doi.org/10.48550/arxiv.2502.16860", "title": "LongAttn: Selecting Long-context Training Data via Token-level Attention", "publication_year": 2025, "publication_date": "2025-02-24", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Wu, Longyun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Longyun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100567405", "display_name": "Dawei Zhu", "orcid": "https://orcid.org/0000-0003-1534-8854", "observed_orcids": ["https://orcid.org/0000-0003-1534-8854"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Dawei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhao, Guangxiang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhao, Guangxiang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5016025493", "display_name": "Zhuocheng Yu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yu, Zhuocheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5022520193", "display_name": "Junfeng Ran", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ran, Junfeng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119181243", "display_name": "Xiangyu Wong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wong, Xiangyu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Sun, Lin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sun, Lin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5058353424", "display_name": "Sujian Li", "orcid": "https://orcid.org/0000-0001-7493-0786", "observed_orcids": ["https://orcid.org/0000-0001-7493-0786"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Sujian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"With": [0], "the": [1, 67, 74, 78, 139], "development": [2], "of": [3, 70, 88], "large": [4], "language": [5], "models": [6], "(LLMs),": [7], "there": [8], "has": [9, 119], "been": [10], "an": [11], "increasing": [12], "need": [13], "for": [14, 77], "significant": [15], "advancements": [16], "in": [17, 50, 131], "handling": [18], "long": [19], "contexts.": [20], "To": [21, 127], "enhance": [22], "long-context": [23, 38, 108, 132, 141], "capabilities,": [24], "constructing": [25], "high-quality": [26, 140], "training": [27, 142], "data": [28, 39, 101, 143], "with": [29], "long-range": [30, 75, 94], "dependencies": [31, 76], "is": [32], "crucial.": [33], "Existing": [34], "methods": [35], "to": [36, 72], "select": [37], "often": [40], "rely": [41], "on": [42], "sentence-level": [43], "analysis,": [44], "which": [45, 65], "can": [46], "be": [47], "greatly": [48], "optimized": [49], "both": [51], "performance": [52], "and": [53, 85, 99, 112, 125, 138], "efficiency.": [54, 126], "In": [55], "this": [56], "paper,": [57], "we": [58, 134], "propose": [59], "a": [60], "novel": [61], "token-level": [62, 82], "framework,": [63], "LongAttn,": [64], "leverages": [66], "self-attention": [68], "mechanism": [69], "LLMs": [71], "measure": [73], "data.": [79], "By": [80], "calculating": [81], "dependency": [83], "strength": [84], "distribution": [86], "uniformity": [87], "token": [89], "scores,": [90], "LongAttn": [91, 118], "effectively": [92], "quantifies": [93], "dependencies,": [95], "enabling": [96], "more": [97], "accurate": [98], "efficient": [100], "selection.": [102], "We": [103], "filter": [104], "LongABC-32K": [105], "from": [106], "open-source": [107], "datasets": [109], "(ArXiv,": [110], "Book,": [111], "Code).": [113], "Through": [114], "our": [115, 136], "comprehensive": [116], "experiments,": [117], "demonstrated": [120], "its": [121], "excellent": [122], "effectiveness,": [123], "scalability,": [124], "facilitate": [128], "future": [129], "research": [130], "data,": [133], "released": [135], "code": [137], "LongABC-32K.": [144]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9779999852180481, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9779999852180481, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.018799999728798866, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.0010999999940395355, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.16860", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.16860", "pdf_url": "https://arxiv.org/pdf/2502.16860", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416716396", "doi": "https://doi.org/10.48550/arxiv.2507.06567", "title": "SlimCaching: Edge Caching of Mixture-of-Experts for Distributed Inference", "publication_year": 2025, "publication_date": "2025-07-09", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100428479", "display_name": "Qian Chen", "orcid": "https://orcid.org/0000-0002-6332-6222", "observed_orcids": ["https://orcid.org/0000-0002-6332-6222"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Qian", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083484070", "display_name": "Xianhao Chen", "orcid": "https://orcid.org/0000-0002-4295-940X", "observed_orcids": ["https://orcid.org/0000-0002-4295-940X", "https://orcid.org/0000-0001-5401-0157", "https://orcid.org/0009-0001-7031-1440"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Xianhao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5007131492", "display_name": "Kaibin Huang", "orcid": "https://orcid.org/0000-0001-8773-4629", "observed_orcids": ["https://orcid.org/0000-0001-8773-4629"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Huang, Kaibin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Mixture-of-Experts": [0], "(MoE)": [1], "models": [2, 9, 190], "improve": [3], "the": [4, 23, 63, 87, 113, 121, 145, 170, 175], "scalability": [5], "of": [6, 17, 26, 151], "large": [7], "language": [8], "(LLMs)": [10], "by": [11, 75, 157], "activating": [12], "only": [13], "a": [14, 34, 48, 71, 91, 103, 107, 138, 149, 158, 179], "small": [15], "subset": [16], "relevant": [18], "experts": [19, 51], "per": [20], "input.": [21], "However,": [22], "sheer": [24], "number": [25], "expert": [27, 66, 77, 118], "networks": [28], "in": [29, 182], "an": [30, 39, 55, 165], "MoE": [31, 123, 189], "model": [32], "introduces": [33, 125], "significant": [35], "storage/memory": [36], "burden": [37], "for": [38, 58, 99], "edge": [40, 56, 80], "device.": [41], "To": [42, 132], "address": [43], "this": [44, 134], "challenge,": [45], "we": [46, 69, 101, 136, 163], "consider": [47], "scenario": [49], "where": [50, 116], "are": [52], "dispersed": [53], "across": [54], "network": [57], "distributed": [59], "inference.": [60], "Based": [61], "on": [62, 79, 169, 187], "popular": [64], "Top-$K$": [65], "selection": [67], "strategy,": [68], "formulate": [70], "latency": [72, 198], "minimization": [73], "problem": [74, 88, 95, 147], "optimizing": [76], "caching": [78], "servers": [81], "under": [82], "storage": [83], "constraints.": [84], "When": [85], "$K=1$,": [86], "reduces": [89, 196], "to": [90, 143, 173, 200], "monotone": [92], "submodular": [93], "maximization": [94], "with": [96, 106, 153, 178], "knapsack": [97], "constraints,": [98], "which": [100, 127], "design": [102, 164], "greedy-based": [104], "algorithm": [105, 167], "$(1": [108], "-": [109], "1/e)$-approximation": [110], "guarantee.": [111], "For": [112], "general": [114], "case": [115], "$K\\geq1$,": [117], "co-activation": [119], "within": [120], "same": [122], "layer": [124], "non-submodularity,": [126], "renders": [128], "greedy": [129, 140], "methods": [130], "ineffective.": [131], "tackle": [133], "issue,": [135], "propose": [137], "successive": [139], "decomposition": [141], "method": [142, 194], "decompose": [144], "original": [146], "into": [148], "series": [150], "subproblems,": [152], "each": [154], "being": [155], "solved": [156], "dynamic": [159], "programming": [160], "approach.": [161], "Furthermore,": [162], "accelerated": [166], "based": [168], "max-convolution": [171], "technique": [172], "obtain": [174], "approximate": [176], "solution": [177], "provable": [180], "guarantee": [181], "polynomial": [183], "time.": [184], "Simulation": [185], "results": [186], "various": [188], "demonstrate": [191], "that": [192], "our": [193], "significantly": [195], "inference": [197], "compared": [199], "existing": [201], "baselines.": [202]}, "primary_topic": {"id": "https://openalex.org/T10273", "display_name": "IoT and Edge/Fog Computing", "score": 0.6675000190734863, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10273", "display_name": "IoT and Edge/Fog Computing", "score": 0.6675000190734863, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11478", "display_name": "Caching and Content Delivery", "score": 0.2741999924182892, "subfield": {"id": "https://openalex.org/subfields/1705", "display_name": "Computer Networks and Communications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T14098", "display_name": "Energy Efficiency in Computing", "score": 0.003700000001117587, "subfield": {"id": "https://openalex.org/subfields/1708", "display_name": "Hardware and Architecture"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2507.06567", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2507.06567", "pdf_url": "https://arxiv.org/pdf/2507.06567", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415787567", "doi": "https://doi.org/10.1038/s41598-025-22198-6", "title": "A Language model-based approach to sentiment classification of languages in central Asia", "publication_year": 2025, "publication_date": "2025-11-03", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5037809353", "display_name": "Palidan Muhetaer", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I94047373", "display_name": "Xinjiang University of Finance and Economics", "ror": "https://ror.org/00fk31757", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I94047373"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Palidan Muhetaer", "raw_affiliation_strings": ["College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China. paridam@aliyun.com", "College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China. paridam@aliyun.com", "institution_ids": ["https://openalex.org/I94047373"]}, {"raw_affiliation_string": "College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China", "institution_ids": ["https://openalex.org/I94047373"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Guo Wenqiang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I94047373", "display_name": "Xinjiang University of Finance and Economics", "ror": "https://ror.org/00fk31757", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I94047373"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Guo Wenqiang", "raw_affiliation_strings": ["College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China", "institution_ids": ["https://openalex.org/I94047373"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5082942072", "display_name": "Chong Lu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I94047373", "display_name": "Xinjiang University of Finance and Economics", "ror": "https://ror.org/00fk31757", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I94047373"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Lu Chong", "raw_affiliation_strings": ["College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "College of Information Management, Xinjiang University of Finance & Economics, Xinjiang, 830012, China", "institution_ids": ["https://openalex.org/I94047373"]}]}], "abstract_inverted_index": {"Xinjiang": [0, 56, 252], "is": [1, 15, 161, 232], "located": [2], "in": [3, 44, 188, 217, 229], "the": [4, 7, 23, 26, 39, 49, 65, 68, 83, 112, 122, 126, 135, 139, 148, 153, 165, 169, 173, 199, 214, 219, 227, 236, 240, 243], "hinterland": [5], "of": [6, 28, 52, 55, 61, 67, 125, 138, 155, 176, 179, 221, 242], "Asia-Europe": [8, 62], "continent,": [9], "neighboring": [10], "Kazakhstan": [11], "and": [12, 14, 25, 41, 58, 73, 78, 87, 100, 134, 158, 168, 205, 213, 254], "Kyrgyzstan,": [13], "a": [16, 35], "bridgehead": [17], "for": [18], "China's": [19], "eastward": [20], "connection": [21], "to": [22, 92, 120, 259], "west,": [24], "classification": [27, 174, 185, 193, 249], "Central": [29, 69, 94, 127, 140, 180, 200, 222, 244], "Asian": [30, 70, 95, 128, 141, 181, 201, 223, 245], "language": [31, 71, 97, 113, 203, 237, 247, 261], "emotions": [32], "has": [33], "become": [34], "development": [36], "focus": [37], "under": [38], "Belt": [40, 253], "Road": [42, 255], "Initiative": [43, 256], "Xinjiang.": [45], "This": [46], "paper": [47, 190, 231], "takes": [48], "Kazakh": [50], "versions": [51], "CHINA": [53], "DAILY": [54], "News": [57, 63], "SILK": [59], "ROAD": [60], "as": [64, 118, 239], "source": [66], "corpus": [72, 77], "constructs": [74], "an": [75], "important": [76], "online": [79], "retrieval": [80], "platform.": [81], "Under": [82], "Transformer": [84], "architecture,": [85], "Word2Vec-TF-IDF": [86], "BERT": [88], "models": [89, 197], "are": [90, 116, 143], "used": [91, 117], "train": [93], "regional": [96, 202, 224, 246], "word": [98, 102, 107], "vectors": [99], "construct": [101], "vector": [103, 108], "features,": [104], "respectively.": [105], "The": [106, 183], "features": [109, 124, 137, 157, 160], "obtained": [110], "from": [111], "pre-training": [114], "model": [115, 186, 228, 238], "inputs": [119], "obtain": [121], "local": [123, 156], "languages": [129, 142, 225], "using": [130, 226], "multi-channel": [131], "convolutional": [132], "CNN,": [133], "global": [136, 159], "extracted": [144], "by": [145], "combining": [146], "with": [147], "bi-directional": [149], "GRU": [150], "model.": [151], "Then": [152], "fusion": [154], "carried": [162], "out": [163], "through": [164], "attention": [166], "mechanism,": [167], "SoftMax": [170], "classifier": [171], "outputs": [172], "results": [175, 194], "sentiment": [177, 184, 220, 248, 262], "tendency": [178], "languages.": [182], "designed": [187], "this": [189, 230], "achieves": [191], "better": [192], "than": [195], "other": [196], "on": [198], "corpus,": [204], "its": [206], "accuracy": [207], "can": [208, 250], "reach": [209], "92.78%,": [210], "91.45%,": [211], "93.54%,": [212], "training": [215], "time": [216], "classifying": [218], "359.71": [233], "s.": [234], "Using": [235], "basis": [241], "help": [251], "implementation": [257], "process": [258], "understand": [260], "changes.": [263]}, "primary_topic": {"id": "https://openalex.org/T10664", "display_name": "Sentiment Analysis and Opinion Mining", "score": 0.9984999895095825, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10664", "display_name": "Sentiment Analysis and Opinion Mining", "score": 0.9984999895095825, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.0010000000474974513, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12262", "display_name": "Hate Speech and Cyberbullying Detection", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1038/s41598-025-22198-6", "is_oa": true, "landing_page_url": "https://doi.org/10.1038/s41598-025-22198-6", "pdf_url": "https://www.nature.com/articles/s41598-025-22198-6.pdf", "source": {"id": "https://openalex.org/S196734849", "display_name": "Scientific Reports", "issn_l": "2045-2322", "issn": ["2045-2322"], "is_oa": true, "is_in_doaj": true, "is_core": true, "listed_in": ["cwts-core", "doaj", "doyens", "jufo-1", "ki-jl-1", "medline", "norway-1"], "host_organization": "https://openalex.org/P4310319908", "host_organization_name": "Nature Portfolio", "host_organization_lineage": ["https://openalex.org/P4310319908", "https://openalex.org/P4310319965"], "host_organization_lineage_names": ["Nature Portfolio", "Springer Nature"], "type": "journal"}, "license": "cc-by-nc-nd", "license_id": "https://openalex.org/licenses/cc-by-nc-nd", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Scientific Reports", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4407632296", "doi": "https://doi.org/10.48550/arxiv.2502.09782", "title": "Improving Acoustic Side-Channel Attacks on Keyboards Using Transformers and Large Language Models", "publication_year": 2025, "publication_date": "2025-02-13", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Park, Jin Hyun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Park, Jin Hyun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5116293071", "display_name": "Seyyed Ali Ayati", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ayati, Seyyed Ali", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Cai, Yichen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cai, Yichen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0], "increasing": [1], "prevalence": [2], "of": [3, 19, 47, 109, 156, 164, 171], "microphones": [4], "in": [5, 129, 180], "everyday": [6], "devices": [7], "and": [8, 36, 45, 76, 92, 125, 158, 177], "the": [9, 17, 43, 58, 96, 107, 161, 168], "growing": [10], "reliance": [11], "on": [12], "online": [13], "services": [14], "have": [15], "amplified": [16], "risk": [18], "acoustic": [20], "side-channel": [21], "attacks": [22], "(ASCAs)": [23], "targeting": [24], "keyboards.": [25], "This": [26, 154], "study": [27], "explores": [28], "deep": [29], "learning": [30], "techniques,": [31], "specifically": [32], "vision": [33], "transformers": [34], "(VTs)": [35], "large": [37], "language": [38, 93, 138], "models": [39, 139, 149], "(LLMs),": [40], "to": [41, 84, 147, 174], "enhance": [42], "effectiveness": [44], "applicability": [46, 163], "such": [48], "attacks.": [49], "We": [50, 87], "present": [51], "substantial": [52], "improvements": [53], "over": [54], "prior": [55], "research,": [56], "with": [57, 95, 140, 150], "CoAtNet": [59, 65], "model": [60, 99], "achieving": [61], "state-of-the-art": [62], "performance.": [63, 102, 134], "Our": [64], "shows": [66], "a": [67, 110], "5.0%": [68], "improvement": [69], "for": [70, 78, 114, 120], "keystrokes": [71, 128], "recorded": [72, 80], "via": [73, 81], "smartphone": [74], "(Phone)": [75], "5.9%": [77], "those": [79], "Zoom": [82], "compared": [83], "previous": [85], "benchmarks.": [86], "also": [88], "evaluate": [89], "transformer": [90], "architectures": [91], "models,": [94], "best": [97], "VT": [98], "matching": [100], "CoAtNet's": [101], "A": [103], "key": [104], "advancement": [105], "is": [106], "introduction": [108], "noise": [111], "mitigation": [112], "method": [113], "real-world": [115, 181], "scenarios.": [116, 182], "By": [117], "using": [118], "LLMs": [119, 159], "contextual": [121], "understanding,": [122], "we": [123], "detect": [124], "correct": [126], "erroneous": [127], "noisy": [130], "environments,": [131], "enhancing": [132], "ASCA": [133, 165], "Additionally,": [135], "fine-tuned": [136], "lightweight": [137], "Low-Rank": [141], "Adaptation": [142], "(LoRA)": [143], "deliver": [144], "comparable": [145], "performance": [146], "heavyweight": [148], "67X": [151], "more": [152], "parameters.": [153], "integration": [155], "VTs": [157], "improves": [160], "practical": [162], "mitigation,": [166], "marking": [167], "first": [169], "use": [170], "these": [172], "technologies": [173], "address": [175], "ASCAs": [176], "error": [178], "correction": [179]}, "primary_topic": {"id": "https://openalex.org/T11800", "display_name": "User Authentication and Security Systems", "score": 0.4578000009059906, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11800", "display_name": "User Authentication and Security Systems", "score": 0.4578000009059906, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.257099986076355, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10951", "display_name": "Cryptographic Implementations and Security", "score": 0.05139999836683273, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.09782", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.09782", "pdf_url": "https://arxiv.org/pdf/2502.09782", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4415145792", "doi": "https://doi.org/10.48550/arxiv.2506.00773", "title": "Dynamic Chunking and Selection for Reading Comprehension of Ultra-Long Context in Large Language Models", "publication_year": 2025, "publication_date": "2025-06-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5110532804", "display_name": "Boheng Sheng", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sheng, Boheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5042407897", "display_name": "Jiacheng Yao", "orcid": "https://orcid.org/0009-0009-5271-0454", "observed_orcids": ["https://orcid.org/0009-0009-5271-0454"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yao, Jiacheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Zhang, Meicong", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhang, Meicong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5000341481", "display_name": "Guoxiu He", "orcid": "https://orcid.org/0000-0002-1419-7495", "observed_orcids": ["https://orcid.org/0000-0002-1419-7495"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "He, Guoxiu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "language": [1], "models": [2], "(LLMs)": [3], "often": [4], "struggle": [5], "to": [6, 36, 79, 93, 137], "accurately": [7], "read": [8], "and": [9, 38, 54, 109, 142], "comprehend": [10], "extremely": [11], "long": [12, 22, 58, 82], "texts.": [13], "Current": [14], "methods": [15], "for": [16, 51, 65, 100], "improvement": [17], "typically": [18], "rely": [19], "on": [20, 106], "splitting": [21], "contexts": [23, 83], "into": [24, 84], "fixed-length": [25], "chunks.": [26, 86], "However,": [27], "fixed": [28], "truncation": [29], "risks": [30], "separating": [31, 53], "semantically": [32], "relevant": [33], "content,": [34], "leading": [35], "ambiguity": [37], "compromising": [39], "accurate": [40], "understanding.": [41], "To": [42], "overcome": [43], "this": [44], "limitation,": [45], "we": [46, 69], "propose": [47], "a": [48, 61, 90, 127], "straightforward": [49], "approach": [50, 117], "dynamically": [52], "selecting": [55], "chunks": [56, 96], "of": [57, 130, 135], "context,": [59], "facilitating": [60], "more": [62], "streamlined": [63], "input": [64, 131], "LLMs.": [66], "In": [67], "particular,": [68], "compute": [70], "semantic": [71], "similarities": [72, 78], "between": [73], "adjacent": [74], "sentences,": [75], "using": [76], "lower": [77], "adaptively": [80], "divide": [81], "variable-length": [85], "We": [87], "further": [88], "train": [89], "question-aware": [91], "classifier": [92], "select": [94], "sensitive": [95], "that": [97, 114], "are": [98, 144], "critical": [99], "answering": [101], "specific": [102], "questions.": [103], "Experimental": [104], "results": [105], "both": [107], "single-hop": [108], "multi-hop": [110], "question-answering": [111], "benchmarks": [112], "show": [113], "the": [115, 147], "proposed": [116], "consistently": [118], "outperforms": [119], "strong": [120], "baselines.": [121], "Notably,": [122], "it": [123], "maintains": [124], "robustness": [125], "across": [126], "wide": [128], "range": [129], "lengths,": [132], "handling": [133], "sequences": [134], "up": [136], "256k": [138], "tokens.": [139], "Our": [140], "datasets": [141], "code": [143], "available": [145], "at": [146], "following": [148], "link:": [149], "https://github.com/ECNU-Text-Computing/DCS": [150]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9897000193595886, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9897000193595886, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.0031999999191612005, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.00279999990016222, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.00773", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.00773", "pdf_url": "https://arxiv.org/pdf/2506.00773", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4408173291", "doi": "https://doi.org/10.1016/j.inffus.2025.103061", "title": "Hyperbolic graph attention network fusing long-context for technical keyphrase extraction", "publication_year": 2025, "publication_date": "2025-03-07", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5015300382", "display_name": "Yushan Zhao", "orcid": "https://orcid.org/0000-0003-1147-4802", "observed_orcids": ["https://orcid.org/0000-0003-1147-4802"]}, "institutions": [{"id": "https://openalex.org/I184681353", "display_name": "Anhui University of Science and Technology", "ror": "https://ror.org/00q9atg80", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I184681353"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yushan Zhao", "raw_affiliation_strings": ["School of Mathematics and Big Data, Anhui University of Science and Technology, Huainan, 232001, Anhui, China"], "raw_orcid": "https://orcid.org/0000-0003-1147-4802", "affiliations": [{"raw_affiliation_string": "School of Mathematics and Big Data, Anhui University of Science and Technology, Huainan, 232001, Anhui, China", "institution_ids": ["https://openalex.org/I184681353"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063188437", "display_name": "Kuan\u2010Ching Li", "orcid": "https://orcid.org/0000-0003-1381-4364", "observed_orcids": ["https://orcid.org/0000-0003-1381-4364"]}, "institutions": [{"id": "https://openalex.org/I177918364", "display_name": "Providence University", "ror": "https://ror.org/03fcpsq87", "country_code": "TW", "type": "education", "lineage": ["https://openalex.org/I177918364"]}], "countries": ["TW"], "is_corresponding": true, "raw_author_name": "Kuan-Ching Li", "raw_affiliation_strings": ["Department of Computer Science and Information Engineering (CSIE), Providence University, Taichung, 43301, Taiwan", "Department of Computer Science and Information Engineering (CSIE), Providence University, Taichung, Taiwan"], "raw_orcid": "https://orcid.org/0000-0003-1381-4364", "affiliations": [{"raw_affiliation_string": "Department of Computer Science and Information Engineering (CSIE), Providence University, Taichung, 43301, Taiwan", "institution_ids": ["https://openalex.org/I177918364"]}, {"raw_affiliation_string": "Department of Computer Science and Information Engineering (CSIE), Providence University, Taichung, Taiwan", "institution_ids": ["https://openalex.org/I177918364"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5061581624", "display_name": "Shunxiang Zhang", "orcid": "https://orcid.org/0000-0002-0540-7593", "observed_orcids": ["https://orcid.org/0000-0002-0540-7593"]}, "institutions": [{"id": "https://openalex.org/I184681353", "display_name": "Anhui University of Science and Technology", "ror": "https://ror.org/00q9atg80", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I184681353"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Shunxiang Zhang", "raw_affiliation_strings": ["School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan, 232001, Anhui, China"], "raw_orcid": "https://orcid.org/0000-0002-0540-7593", "affiliations": [{"raw_affiliation_string": "School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan, 232001, Anhui, China", "institution_ids": ["https://openalex.org/I184681353"]}]}, {"author_position": "last", "author": {"id": null, "display_name": "Tongzhou Ye", "orcid": "https://orcid.org/0000-0002-0249-4353", "observed_orcids": ["https://orcid.org/0000-0002-0249-4353"]}, "institutions": [{"id": "https://openalex.org/I184681353", "display_name": "Anhui University of Science and Technology", "ror": "https://ror.org/00q9atg80", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I184681353"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Tongzhou Ye", "raw_affiliation_strings": ["School of Mechanical and Electrical Engineering, Anhui University of Science and Technology, Huainan, 232001, Anhui, China"], "raw_orcid": "https://orcid.org/0000-0002-0249-4353", "affiliations": [{"raw_affiliation_string": "School of Mechanical and Electrical Engineering, Anhui University of Science and Technology, Huainan, 232001, Anhui, China", "institution_ids": ["https://openalex.org/I184681353"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T13083", "display_name": "Advanced Text Analysis Techniques", "score": 0.9976000189781189, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T13083", "display_name": "Advanced Text Analysis Techniques", "score": 0.9976000189781189, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.002099999925121665, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12881", "display_name": "linguistics and terminology studies", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1203", "display_name": "Language and Linguistics"}, "field": {"id": "https://openalex.org/fields/12", "display_name": "Arts and Humanities"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "doi:10.1016/j.inffus.2025.103061", "is_oa": false, "landing_page_url": "https://doi.org/10.1016/j.inffus.2025.103061", "pdf_url": null, "source": {"id": "https://openalex.org/S7560371", "display_name": "Information Fusion", "issn_l": "1566-2535", "issn": ["1566-2535", "1872-6305"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-3", "norway-1"], "host_organization": "https://openalex.org/P4310320990", "host_organization_name": "Elsevier BV", "host_organization_lineage": ["https://openalex.org/P4310320990"], "host_organization_lineage_names": ["Elsevier BV"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Information Fusion", "raw_type": "journal-article"}, "cited_by_count": 4, "is_retracted": false}, {"id": "https://openalex.org/W4416054669", "doi": "https://doi.org/10.48550/arxiv.2510.17934", "title": "AtlasKV: Augmenting LLMs with Billion-Scale Knowledge Graphs in 20GB VRAM", "publication_year": 2025, "publication_date": "2025-10-20", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Huang, Haoyu", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I200769079", "display_name": "Hong Kong University of Science and Technology", "ror": "https://ror.org/00q4vv597", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I200769079"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "Huang, Haoyu", "raw_affiliation_strings": ["The Hong Kong University of Science and Technology"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Hong Kong University of Science and Technology", "institution_ids": ["https://openalex.org/I200769079"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Tsang, Hong Ting", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I200769079", "display_name": "Hong Kong University of Science and Technology", "ror": "https://ror.org/00q4vv597", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I200769079"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "Tsang, Hong Ting", "raw_affiliation_strings": ["The Hong Kong University of Science and Technology"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Hong Kong University of Science and Technology", "institution_ids": ["https://openalex.org/I200769079"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Bai, Jiaxin", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I200769079", "display_name": "Hong Kong University of Science and Technology", "ror": "https://ror.org/00q4vv597", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I200769079"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "Bai, Jiaxin", "raw_affiliation_strings": ["The Hong Kong University of Science and Technology"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Hong Kong University of Science and Technology", "institution_ids": ["https://openalex.org/I200769079"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Peng, Xi", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I2250955327", "display_name": "Huawei Technologies (China)", "ror": "https://ror.org/00cmhce21", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I2250955327"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Peng, Xi", "raw_affiliation_strings": ["Theory Lab , Huawei"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Theory Lab , Huawei", "institution_ids": ["https://openalex.org/I2250955327"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110719819", "display_name": "Gong Zhang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I2250955327", "display_name": "Huawei Technologies (China)", "ror": "https://ror.org/00cmhce21", "country_code": "CN", "type": "company", "lineage": ["https://openalex.org/I2250955327"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhang, Gong", "raw_affiliation_strings": ["Theory Lab , Huawei"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Theory Lab , Huawei", "institution_ids": ["https://openalex.org/I2250955327"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5020880385", "display_name": "Yangqiu Song", "orcid": "https://orcid.org/0000-0002-7818-6090", "observed_orcids": ["https://orcid.org/0000-0002-7818-6090"]}, "institutions": [{"id": "https://openalex.org/I200769079", "display_name": "Hong Kong University of Science and Technology", "ror": "https://ror.org/00q4vv597", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I200769079"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "Song, Yangqiu", "raw_affiliation_strings": ["The Hong Kong University of Science and Technology"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The Hong Kong University of Science and Technology", "institution_ids": ["https://openalex.org/I200769079"]}]}], "abstract_inverted_index": {"Retrieval-augmented": [0], "generation": [1], "(RAG)": [2], "has": [3], "shown": [4], "some": [5], "success": [6], "in": [7], "augmenting": [8], "large": [9, 42], "language": [10], "models": [11], "(LLMs)": [12], "with": [13, 82, 116], "external": [14, 30, 139], "knowledge.": [15, 150], "However,": [16], "as": [17], "a": [18, 66, 73], "non-parametric": [19], "knowledge": [20, 44, 68, 84, 125], "integration": [21, 69], "paradigm": [22], "for": [23, 40], "LLMs,": [24], "RAG": [25], "methods": [26], "heavily": [27], "rely": [28], "on": [29], "retrieval": [31], "modules": [32], "and": [33, 56, 76, 106, 119, 127, 136], "the": [34, 131], "retrieved": [35], "textual": [36], "context": [37, 142], "prior.": [38], "Especially": [39], "very": [41, 91], "scale": [43, 115], "augmentation,": [45], "they": [46], "would": [47], "introduce": [48, 104], "substantial": [49], "inference": [50], "latency": [51], "due": [52], "to": [53, 79, 108, 148], "expensive": [54], "searches": [55], "much": [57], "longer": [58], "relevant": [59], "context.": [60], "In": [61, 101], "this": [62], "paper,": [63], "we": [64, 103], "propose": [65], "parametric": [67], "method,": [70], "called": [71], "\\textbf{AtlasKV},": [72], "scalable,": [74], "effective,": [75], "general": [77], "way": [78], "augment": [80], "LLMs": [81, 113], "billion-scale": [83], "graphs": [85], "(KGs)": [86], "(e.g.": [87, 96], "1B": [88], "triples)": [89], "using": [90, 130], "little": [92], "GPU": [93], "memory": [94, 120], "cost": [95], "less": [97], "than": [98], "20GB": [99], "VRAM).": [100], "AtlasKV,": [102], "KG2KV": [105], "HiKVP": [107], "integrate": [109], "KG": [110], "triples": [111], "into": [112], "at": [114], "sub-linear": [117], "time": [118], "complexity.": [121], "It": [122], "maintains": [123], "strong": [124], "grounding": [126], "generalization": [128], "performance": [129], "LLMs'": [132], "inherent": [133], "attention": [134], "mechanism,": [135], "requires": [137], "no": [138], "retrievers,": [140], "long": [141], "priors,": [143], "or": [144], "retraining": [145], "when": [146], "adapting": [147], "new": [149]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9506999850273132, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9506999850273132, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.017400000244379044, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11273", "display_name": "Advanced Graph Neural Networks", "score": 0.016699999570846558, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.17934", "is_oa": true, "landing_page_url": "https://arxiv.org/abs/2510.17934", "pdf_url": "https://arxiv.org/pdf/2510.17934", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7119156389", "doi": "https://doi.org/10.34734/fzj-2025-05839", "title": "Earth Action in Transition: Highlights From the 2025 ESA\u2013NASA International Workshop on AI Foundation Models for EO [Space-Agencies]", "publication_year": 2025, "publication_date": "2025-01-01", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5009020295", "display_name": "Nicolas Long\u00e9p\u00e9", "orcid": "https://orcid.org/0000-0002-6832-3274", "observed_orcids": ["https://orcid.org/0000-0002-6832-3274"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Long\u00e9p\u00e9, Nicolas", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0002-6832-3274", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5122062008", "display_name": "Hamed Alemohammad", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Alemohammad, Hamed", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-5662-3643", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5051374473", "display_name": "Anca Anghelea", "orcid": "https://orcid.org/0000-0002-8503-7608", "observed_orcids": ["https://orcid.org/0000-0002-8503-7608"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Anghelea, Anca", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0002-8503-7608", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5034679430", "display_name": "Thomas Brunschwiler", "orcid": "https://orcid.org/0000-0002-7254-3405", "observed_orcids": ["https://orcid.org/0000-0002-7254-3405"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Brunschwiler, Thomas", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0002-7254-3405", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5039052506", "display_name": "Gustau Camps\u2010Valls", "orcid": "https://orcid.org/0000-0003-1683-2138", "observed_orcids": ["https://orcid.org/0000-0003-1683-2138"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Camps-Valls, Gustau", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0003-1683-2138", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5058680339", "display_name": "Gabriele Cavallaro", "orcid": "https://orcid.org/0000-0002-3239-9904", "observed_orcids": ["https://orcid.org/0000-0002-3239-9904"]}, "institutions": [{"id": "https://openalex.org/I171892758", "display_name": "Forschungszentrum J\u00fclich", "ror": "https://ror.org/02nv7yv05", "country_code": "DE", "type": "facility", "lineage": ["https://openalex.org/I1305996414", "https://openalex.org/I171892758"]}], "countries": ["DE"], "is_corresponding": false, "raw_author_name": "Cavallaro, Gabriele", "raw_affiliation_strings": ["Forschungszentrum J\u00fclich"], "raw_orcid": "https://orcid.org/0000-0002-3239-9904", "affiliations": [{"raw_affiliation_string": "Forschungszentrum J\u00fclich", "institution_ids": ["https://openalex.org/I171892758"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5109059550", "display_name": "J. Chanussot", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chanussot, Jocelyn", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0003-4817-2875", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5122281268", "display_name": "Jose Manuel Delgado", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Delgado, Jose Manuel", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5126500644", "display_name": "Beg\u00fcm Demir", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Demir, Beg\u00fcm", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0003-2175-7072", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5052528008", "display_name": "Nikolaos Dionelis", "orcid": "https://orcid.org/0000-0001-9662-8537", "observed_orcids": ["https://orcid.org/0000-0001-9662-8537"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dionelis, Nikolaos", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-9662-8537", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5062702059", "display_name": "Paolo Fraccaro", "orcid": "https://orcid.org/0000-0002-5213-7071", "observed_orcids": ["https://orcid.org/0000-0002-5213-7071"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Fraccaro, Paolo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066516763", "display_name": "Anna Jungbluth", "orcid": "https://orcid.org/0000-0002-9888-6262", "observed_orcids": ["https://orcid.org/0000-0002-9888-6262"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jungbluth, Anna", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5030258636", "display_name": "Robert E. Kennedy", "orcid": "https://orcid.org/0000-0002-5507-474X", "observed_orcids": ["https://orcid.org/0000-0002-5507-474X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kennedy, Robert E.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5089472869", "display_name": "Valerio Marsocci", "orcid": "https://orcid.org/0000-0003-1442-7155", "observed_orcids": ["https://orcid.org/0000-0003-1442-7155"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Marsocci, Valerio", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5024141403", "display_name": "Muthukumaran Ramasubramanian", "orcid": "https://orcid.org/0000-0001-5293-8349", "observed_orcids": ["https://orcid.org/0000-0001-5293-8349"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ramasubramanian, Muthukumaran", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-5293-8349", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5122190107", "display_name": "Raul Ramos-Pollan", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ramos-Pollan, Raul", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-6195-3612", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5122227751", "display_name": "Sujit Roy", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Roy, Sujit", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5014535551", "display_name": "Gencer S\u00fcmb\u00fcl", "orcid": "https://orcid.org/0000-0001-5343-7992", "observed_orcids": ["https://orcid.org/0000-0001-5343-7992"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "S\u00fcmb\u00fcl, Gencer", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5005192117", "display_name": "Devis Tuia", "orcid": "https://orcid.org/0000-0003-0374-2459", "observed_orcids": ["https://orcid.org/0000-0003-0374-2459"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tuia, Devis", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0003-0374-2459", "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5122270737", "display_name": "Xiao Xiang Zhu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Xiao Xiang", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0001-8107-9096", "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5010966122", "display_name": "Rahul Ramachandran", "orcid": "https://orcid.org/0000-0002-0647-1941", "observed_orcids": ["https://orcid.org/0000-0002-0647-1941"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ramachandran, Rahul", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0000-0002-0647-1941", "affiliations": []}], "abstract_inverted_index": {"Over": [0], "850": [1], "people": [2, 38], "joined": [3], "the": [4, 33, 47, 65, 106, 209, 246], "first": [5], "International": [6], "Workshop": [7], "on": [8, 39, 232], "AI": [9, 101, 213], "Foundation": [10], "Model": [11], "(FM)": [12], "for": [13, 108, 146, 180], "Earth": [14, 29, 55], "Observation": [15, 30], "(EO)": [16], "co-organized": [17], "by": [18, 220], "ESA": [19], "and": [20, 41, 54, 67, 77, 86, 94, 110, 115, 126, 152, 157, 166, 174, 223, 230, 255], "NASA": [21], "5-7": [22], "May": [23], "2025.": [24], "Hosted": [25], "at": [26], "ESRIN": [27], "(ESA\u2019s": [28], "Center,": [31], "Italy),": [32], "event": [34], "welcomed": [35], "around": [36], "300": [37], "site,": [40], "an": [42], "additional": [43], "550": [44], "online,": [45], "with": [46, 141, 216, 257], "promise": [48], "that": [49, 80, 226], "FMs": [50, 142, 185, 222], "can": [51, 227], "revolutionize": [52], "EO": [53, 66, 89, 182, 221, 233], "sciences.": [56], "The": [57, 103, 132, 192, 206], "workshop": [58, 104, 207], "marked": [59], "a": [60, 71], "pivotal": [61], "moment": [62], "in": [63, 214], "aligning": [64], "FM": [68], "communities,": [69], "fostering": [70], "shared": [72], "commitment": [73], "to": [74, 122, 127, 138, 162, 196, 240], "developing": [75], "open": [76], "trustworthy": [78], "tools": [79], "support": [81], "science": [82], "discovery,": [83], "operational": [84], "applications,": [85], "prescriptive": [87], "analytics.": [88, 183], "data": [90], "is": [91], "massive,": [92], "complex": [93], "high": [95], "dimensional": [96], "requiring": [97], "specific": [98], "yet": [99], "scalable": [100], "architectures.": [102], "emphasized": [105], "need": [107], "training": [109], "architecture": [111], "enabling": [112], "interpretability,": [113], "explainability,": [114], "physical": [116], "consistency.": [117], "Coordination": [118], "should": [119], "be": [120], "strengthened": [121], "minimize": [123], "redundant": [124], "development": [125, 165], "better": [128], "leverage": [129], "collective": [130], "expertise.": [131], "focus": [133], "has": [134], "shifted": [135], "from": [136, 237], "prototyping": [137], "real-world": [139], "deployment,": [140], "needing": [143], "further": [144], "design": [145], "integration": [147], "into": [148], "digital": [149], "twins,": [150], "dashboards,": [151], "edge": [153], "platforms.": [154], "Transparent": [155], "benchmarking": [156], "user-driven": [158], "evaluation": [159], "are": [160], "key": [161, 253], "guiding": [163], "model": [164], "decision-making.": [167], "In": [168], "addition,": [169], "parameter-efficient": [170], "adaptation,": [171], "neural": [172], "compression,": [173], "embedding-based": [175], "workflows": [176], "offer": [177], "promising": [178], "paths": [179], "scaling": [181], "While": [184], "show": [186], "promise,": [187], "their": [188], "effectiveness": [189], "remains": [190], "context-dependent.": [191], "community": [193], "debated": [194], "whether": [195], "pursue": [197], "universal": [198], "models,": [199, 225], "specialized": [200], "solutions,": [201], "or": [202], "mixtures": [203], "of": [204, 211, 248], "experts.": [205], "envisioned": [208], "future": [210, 247], "agentic": [212], "EO,": [215], "multi-agent": [217], "system": [218], "powered": [219], "vision-language": [224], "dynamically": [228], "reason": [229], "act": [231], "data.": [234], "This": [235, 250], "shift": [236], "static": [238], "pipelines": [239], "adaptive,": [241], "smarter": [242], "systems": [243], "could": [244], "redefine": [245], "EO.": [249], "paper": [251], "summarizes": [252], "discussions": [254], "concludes": [256], "thought-provoking": [258], "remarks.": [259]}, "primary_topic": {"id": "https://openalex.org/T10689", "display_name": "Remote-Sensing Image Classification", "score": 0.5214999914169312, "subfield": {"id": "https://openalex.org/subfields/2214", "display_name": "Media Technology"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10689", "display_name": "Remote-Sensing Image Classification", "score": 0.5214999914169312, "subfield": {"id": "https://openalex.org/subfields/2214", "display_name": "Media Technology"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.09929999709129333, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0731000006198883, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.34734/fzj-2025-05839", "is_oa": true, "landing_page_url": "https://doi.org/10.34734/fzj-2025-05839", "pdf_url": null, "source": {"id": "https://openalex.org/S2491948244", "display_name": "IEEE Geoscience and Remote Sensing Magazine", "issn_l": "2168-6831", "issn": ["2168-6831", "2373-7468", "2473-2397"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-2"], "host_organization": "https://openalex.org/P4310319808", "host_organization_name": "Institute of Electrical and Electronics Engineers", "host_organization_lineage": ["https://openalex.org/P4310319808"], "host_organization_lineage_names": ["Institute of Electrical and Electronics Engineers"], "type": "journal"}, "license": null, "license_id": null, "version": null, "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "article-journal"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4412513184", "doi": "https://doi.org/10.2316/j.2025.206-1233", "title": "MULTIMODAL INFORMATION PROCESSING AND BEHAVIOUR GENERATION OF HUMANOID ROBOTS BASED ON PALM 2 MODEL AND MULTIMODAL TRANSFORMER ARCHITECTURE, 648-664. SI", "publication_year": 2025, "publication_date": "2025-01-01", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Jie Fang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jie Fang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Wenliang Ju", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wenliang Ju", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Feng Xiao", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Feng Xiao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Rubing Huang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Rubing Huang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Yanjun Liu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Yanjun Liu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"In": [0], "response": [1], "to": [2, 35], "the": [3], "low": [4], "e\ufb03ciency": [5], "of": [6], "behaviour": [7, 41], "decision-makingand": [8], "generation": [9, 42], "for": [10], "humanoid": [11], "robots,": [12], "single": [13], "data": [14], "modality,": [15], "anddi\ufb03culty": [16], "in": [17], "e\ufb00ectively": [18], "integrating": [19], "information": [20, 38], "from": [21], "di\ufb00erentmodalities,": [22], "this": [23], "article": [24], "combines": [25], "Pathways": [26], "Language": [27], "Model": [28], "2(PaLM": [29], "2)": [30], "and": [31, 40], "multimodal": [32], "Transformer": [33], "architecture": [34], "study": [36], "themultimodal": [37], "processing": [39], "ofhumanoid": [43], "robots.": [44]}, "primary_topic": {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.7631999850273132, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.7631999850273132, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10653", "display_name": "Robot Manipulation and Learning", "score": 0.09749999642372131, "subfield": {"id": "https://openalex.org/subfields/2207", "display_name": "Control and Systems Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12031", "display_name": "Speech and dialogue systems", "score": 0.041099999099969864, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2316/j.2025.206-1233", "is_oa": false, "landing_page_url": "https://doi.org/10.2316/j.2025.206-1233", "pdf_url": null, "source": {"id": "https://openalex.org/S74998426", "display_name": "International Journal of Robotics and Automation", "issn_l": "0826-8185", "issn": ["0826-8185", "1925-7090"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1"], "host_organization": null, "host_organization_name": null, "host_organization_lineage": [], "host_organization_lineage_names": [], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "International Journal of Robotics and Automation", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7117174144", "doi": null, "title": "MixKVQ: Query-Aware Mixed-Precision KV Cache Quantization for Long-Context Reasoning", "publication_year": 2025, "publication_date": "2025-12-22", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5121175017", "display_name": "Tao Zhang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I90610280", "display_name": "South China University of Technology", "ror": "https://ror.org/0530pts50", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I90610280"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhang, Tao", "raw_affiliation_strings": ["South China University of Technology , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "South China University of Technology , China,", "institution_ids": ["https://openalex.org/I90610280"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101403806", "display_name": "Ziqian Zeng", "orcid": "https://orcid.org/0000-0003-0060-7956", "observed_orcids": ["https://orcid.org/0000-0003-0060-7956"]}, "institutions": [{"id": "https://openalex.org/I90610280", "display_name": "South China University of Technology", "ror": "https://ror.org/0530pts50", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I90610280"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zeng, Ziqian", "raw_affiliation_strings": ["South China University of Technology , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "South China University of Technology , China,", "institution_ids": ["https://openalex.org/I90610280"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5121131136", "display_name": "Hao Peng", "orcid": "https://orcid.org/0009-0006-5203-5936", "observed_orcids": ["https://orcid.org/0009-0006-5203-5936"]}, "institutions": [{"id": "https://openalex.org/I82880672", "display_name": "Beihang University", "ror": "https://ror.org/00wk2mp56", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I82880672"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Peng, Hao", "raw_affiliation_strings": ["Beihang University , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beihang University , China,", "institution_ids": ["https://openalex.org/I82880672"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5121162302", "display_name": "Huiping Zhuang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I90610280", "display_name": "South China University of Technology", "ror": "https://ror.org/0530pts50", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I90610280"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhuang, Huiping", "raw_affiliation_strings": ["South China University of Technology , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "South China University of Technology , China,", "institution_ids": ["https://openalex.org/I90610280"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5121167888", "display_name": "Cen Chen", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I90610280", "display_name": "South China University of Technology", "ror": "https://ror.org/0530pts50", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I90610280"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chen, Cen", "raw_affiliation_strings": ["Pazhou Laboratory , China ,", "South China University of Technology , China,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Pazhou Laboratory , China ,", "institution_ids": []}, {"raw_affiliation_string": "South China University of Technology , China,", "institution_ids": ["https://openalex.org/I90610280"]}]}], "abstract_inverted_index": {"Long": [0], "Chain-of-Thought": [1], "(CoT)": [2], "reasoning": [3, 51, 140], "has": [4], "significantly": [5, 146], "advanced": [6], "the": [7, 26, 61, 100], "capabilities": [8], "of": [9], "Large": [10], "Language": [11], "Models": [12], "(LLMs),": [13], "but": [14], "this": [15, 104], "progress": [16], "is": [17, 35], "accompanied": [18], "by": [19], "substantial": [20], "memory": [21, 162], "and": [22, 96, 121], "latency": [23], "overhead": [24], "from": [25], "extensive": [27], "Key-Value": [28], "(KV)": [29], "cache.": [30, 136], "Although": [31], "KV": [32, 82], "cache": [33, 83], "quantization": [34, 42, 54, 84, 94, 133], "a": [36, 90, 109, 115, 155, 159], "promising": [37], "compression": [38], "technique,": [39], "existing": [40, 148], "low-bit": [41, 81, 149], "methods": [43], "often": [44], "exhibit": [45], "severe": [46], "performance": [47, 152], "degradation": [48], "on": [49, 103, 138], "complex": [50, 139], "tasks.": [52], "Fixed-precision": [53], "struggles": [55], "to": [56, 69, 99, 119, 154], "handle": [57], "outlier": [58], "channels": [59, 125], "in": [60], "key": [62, 91, 124], "cache,": [63], "while": [64, 130], "current": [65], "mixed-precision": [66], "strategies": [67], "fail": [68], "accurately": [70], "identify": [71, 120], "components": [72], "requiring": [73], "high-precision": [74], "representation.": [75], "We": [76], "find": [77], "that": [78, 113, 126, 143], "an": [79], "effective": [80], "strategy": [85], "must": [86], "consider": [87], "two": [88], "factors:": [89], "channel's": [92], "intrinsic": [93], "difficulty": [95], "its": [97], "relevance": [98], "query.": [101], "Based": [102], "insight,": [105], "we": [106], "propose": [107], "MixKVQ,": [108], "novel": [110], "plug-and-play": [111], "method": [112], "introduces": [114], "lightweight,": [116], "query-aware": [117], "algorithm": [118], "preserve": [122], "critical": [123], "need": [127], "higher": [128], "precision,": [129], "applying": [131], "per-token": [132], "for": [134], "value": [135], "Experiments": [137], "datasets": [141], "demonstrate": [142], "our": [144], "approach": [145], "outperforms": [147], "methods,": [150], "achieving": [151], "comparable": [153], "full-precision": [156], "baseline": [157], "at": [158], "substantially": [160], "reduced": [161], "footprint.": [163]}, "primary_topic": {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.8672000169754028, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.8672000169754028, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.0478999987244606, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.012900000438094139, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.19206", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.19206", "pdf_url": "https://arxiv.org/pdf/2512.19206", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W7115596576", "doi": "https://doi.org/10.48550/arxiv.2512.11221", "title": "Adaptive Soft Rolling KV Freeze with Entropy-Guided Recovery: Sublinear Memory Growth for Efficient LLM Inference", "publication_year": 2025, "publication_date": "2025-12-12", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5130364278", "display_name": "Adilet Metinov", "orcid": "https://orcid.org/0009-0007-6400-4173", "observed_orcids": ["https://orcid.org/0009-0007-6400-4173"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Metinov, Adilet", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Kudakeeva, Gulida M.", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kudakeeva, Gulida M.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Nursultan, Bolotbek uulu", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Nursultan, Bolotbek uulu", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "Kabaeva, Gulnara D.", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kabaeva, Gulnara D.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0, 62], "present": [1], "Adaptive": [2], "Soft": [3], "Rolling": [4], "KV": [5, 92], "Freeze": [6], "with": [7, 66, 75], "Entropy-Guided": [8], "Recovery": [9], "(ASR-KF-EGR),": [10], "a": [11, 24, 39, 113], "training-free": [12], "inference-time": [13], "framework": [14, 65], "for": [15, 34, 116], "efficient": [16], "large": [17], "language": [18], "model": [19], "generation.": [20], "Our": [21], "method": [22, 105], "introduces": [23], "reversible": [25], "soft-freeze": [26], "mechanism": [27], "that": [28, 46], "temporarily": [29], "suspends": [30], "key-value": [31], "(KV)": [32], "updates": [33], "low-importance": [35, 77], "tokens": [36, 53], "identified": [37], "within": [38], "sliding": [40], "attention": [41], "window.": [42], "Unlike": [43], "eviction-based": [44], "approaches": [45], "permanently": [47], "discard": [48], "context,": [49], "ASR-KF-EGR": [50], "preserves": [51], "all": [52], "in": [54, 90], "off-GPU": [55], "storage": [56], "and": [57, 99, 111], "restores": [58], "them": [59], "on": [60, 84], "demand.": [61], "extend": [63], "the": [64], "sublinear": [67], "freeze": [68, 71], "scheduling,": [69], "where": [70], "duration": [72], "grows": [73], "sublinearly": [74], "repeated": [76], "detections,": [78], "preventing": [79], "over-aggressive": [80], "compression.": [81], "Preliminary": [82], "experiments": [83], "LLaMA-3": [85], "8B": [86], "demonstrate": [87], "55-67%": [88], "reduction": [89], "active": [91], "cache": [93], "size": [94], "while": [95], "maintaining": [96], "generation": [97], "quality": [98], "passing": [100], "needle-in-haystack": [101], "retrieval": [102], "tests.": [103], "The": [104], "is": [106], "architecture-agnostic,": [107], "requires": [108], "no": [109], "fine-tuning,": [110], "provides": [112], "practical": [114], "solution": [115], "memory-constrained": [117], "deployment": [118], "of": [119], "long-context": [120], "LLMs.": [121]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5059000253677368, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5059000253677368, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.3792000114917755, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.034699998795986176, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.11221", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.11221", "pdf_url": "https://arxiv.org/pdf/2512.11221", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414634795", "doi": "https://doi.org/10.48550/arxiv.2504.16083", "title": "MMInference: Accelerating Pre-filling for Long-Context VLMs via Modality-Aware Permutation Sparse Attention", "publication_year": 2025, "publication_date": "2025-04-22", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100381401", "display_name": "Yucheng Li", "orcid": "https://orcid.org/0000-0003-3473-3176", "observed_orcids": ["https://orcid.org/0000-0003-3473-3176"]}, "institutions": [{"id": "https://openalex.org/I28290843", "display_name": "University of Surrey", "ror": "https://ror.org/00ks66431", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I28290843"]}], "countries": ["GB"], "is_corresponding": false, "raw_author_name": "Li, Yucheng", "raw_affiliation_strings": ["University of Surrey"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Surrey", "institution_ids": ["https://openalex.org/I28290843"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070156365", "display_name": "Huiqiang Jiang", "orcid": "https://orcid.org/0000-0002-1327-4882", "observed_orcids": ["https://orcid.org/0000-0002-1327-4882"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Jiang, Huiqiang", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5040430995", "display_name": "Chengruidong Zhang", "orcid": "https://orcid.org/0009-0002-4939-0436", "observed_orcids": ["https://orcid.org/0009-0002-4939-0436"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Zhang, Chengruidong", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5024011052", "display_name": "Qianhui Wu", "orcid": "https://orcid.org/0000-0001-9146-0675", "observed_orcids": ["https://orcid.org/0000-0001-9146-0675"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Wu, Qianhui", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100929082", "display_name": "Xufang Luo", "orcid": "https://orcid.org/0000-0002-3405-554X", "observed_orcids": ["https://orcid.org/0000-0002-3405-554X"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Luo, Xufang", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5083787197", "display_name": "Surin Ahn", "orcid": "https://orcid.org/0000-0001-5615-4859", "observed_orcids": ["https://orcid.org/0000-0001-5615-4859"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Ahn, Surin", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Abdi, Amir H.", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Abdi, Amir H.", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100440904", "display_name": "Dongsheng Li", "orcid": "https://orcid.org/0000-0001-9962-1373", "observed_orcids": ["https://orcid.org/0000-0001-9962-1373"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Li, Dongsheng", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5091035881", "display_name": "Jianfeng Gao", "orcid": "https://orcid.org/0000-0001-5540-7228", "observed_orcids": ["https://orcid.org/0000-0001-5540-7228"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Gao, Jianfeng", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5118093092", "display_name": "Yuqing Yang", "orcid": "https://orcid.org/0000-0001-8182-1350", "observed_orcids": ["https://orcid.org/0000-0001-8182-1350"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Yang, Yuqing", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5032760128", "display_name": "Lili Qiu", "orcid": "https://orcid.org/0000-0002-1590-9749", "observed_orcids": ["https://orcid.org/0000-0002-1590-9749", "https://orcid.org/0009-0003-8131-7439"]}, "institutions": [{"id": "https://openalex.org/I4210105678", "display_name": "Microsoft (Finland)", "ror": "https://ror.org/01nehjf29", "country_code": "FI", "type": "company", "lineage": ["https://openalex.org/I1290206253", "https://openalex.org/I4210105678"]}], "countries": ["FI"], "is_corresponding": false, "raw_author_name": "Qiu, Lili", "raw_affiliation_strings": ["Microsoft Corporation "], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Microsoft Corporation ", "institution_ids": ["https://openalex.org/I4210105678"]}]}], "abstract_inverted_index": {"The": [0], "integration": [1], "of": [2, 67], "long-context": [3, 54, 161], "capabilities": [4], "with": [5], "visual": [6], "understanding": [7], "unlocks": [8], "unprecedented": [9], "potential": [10], "for": [11, 53, 112, 131], "Vision": [12], "Language": [13], "Models": [14], "(VLMs).": [15], "However,": [16], "the": [17, 22, 50, 62, 76, 96, 108, 117, 123, 171], "quadratic": [18], "attention": [19, 46], "complexity": [20], "during": [21], "pre-filling": [23, 172], "phase": [24], "remains": [25], "a": [26, 43, 72, 91], "significant": [27], "obstacle": [28], "to": [29, 71, 94, 176], "real-world": [30], "deployment.": [31], "To": [32], "overcome": [33], "this": [34], "limitation,": [35], "we": [36], "introduce": [37, 90], "MMInference": [38, 115, 136, 169], "(Multimodality": [39], "Million": [40], "tokens": [41, 180], "Inference),": [42], "dynamic": [44], "sparse": [45, 74, 84, 110, 118, 133], "method": [47, 93], "that": [48, 61, 168], "accelerates": [49, 170], "prefilling": [51], "stage": [52, 173], "multi-modal": [55, 151], "inputs.": [56], "First,": [57], "our": [58], "analysis": [59], "reveals": [60], "temporal": [63], "and": [64, 100, 157], "spatial": [65], "locality": [66], "video": [68], "input": [69], "leads": [70], "unique": [73, 97], "pattern,": [75], "Grid": [77, 98], "pattern.": [78], "Simultaneously,": [79], "VLMs": [80, 162], "exhibit": [81], "markedly": [82], "different": [83, 87], "distributions": [85], "across": [86], "modalities.": [88], "We": [89, 125], "permutation-based": [92], "leverage": [95], "pattern": [99], "handle": [101], "modality": [102], "boundary": [103], "issues.": [104], "By": [105], "offline": [106], "search": [107], "optimal": [109], "patterns": [111], "each": [113], "head,": [114], "constructs": [116], "distribution": [119], "dynamically": [120], "based": [121], "on": [122, 150], "input.": [124], "also": [126], "provide": [127], "optimized": [128], "GPU": [129], "kernels": [130], "efficient": [132], "computations.": [134], "Notably,": [135], "integrates": [137], "seamlessly": [138], "into": [139], "existing": [140], "VLM": [141], "pipelines": [142], "without": [143], "any": [144], "model": [145], "modifications": [146], "or": [147], "fine-tuning.": [148], "Experiments": [149], "benchmarks-including": [152], "Video": [153], "QA,": [154], "Captioning,": [155], "VisionNIAH,": [156], "Mixed-Modality": [158], "NIAH-with": [159], "state-of-the-art": [160], "(LongVila,": [163], "LlavaVideo,": [164], "VideoChat-Flash,": [165], "Qwen2.5-VL)": [166], "show": [167], "by": [174], "up": [175], "8.3x": [177], "at": [178, 188], "1M": [179], "while": [181], "maintaining": [182], "accuracy.": [183], "Our": [184], "code": [185], "is": [186], "available": [187], "https://aka.ms/MMInference.": [189]}, "primary_topic": {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.9984999895095825, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.9984999895095825, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0008999999845400453, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11605", "display_name": "Visual Attention and Saliency Detection", "score": 9.999999747378752e-05, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2504.16083", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2504.16083", "pdf_url": "https://arxiv.org/pdf/2504.16083", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416047623", "doi": "https://doi.org/10.48550/arxiv.2505.21918", "title": "Self-supervised Learning Method Using Transformer for Multi-dimensional Sensor Data Processing", "publication_year": 2025, "publication_date": "2025-05-28", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5008566358", "display_name": "Kai Xiong Hu", "orcid": "https://orcid.org/0000-0001-8829-1700", "observed_orcids": ["https://orcid.org/0000-0001-8829-1700"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Kai, Haruki", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5026921207", "display_name": "Tsuyoshi Okita", "orcid": "https://orcid.org/0000-0002-1286-5496", "observed_orcids": ["https://orcid.org/0000-0002-1286-5496"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Okita, Tsuyoshi", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0, 96], "developed": [1], "a": [2, 20, 62, 83, 89], "deep": [3], "learning": [4], "algorithm": [5], "for": [6], "human": [7, 52], "activity": [8, 53], "recognition": [9], "using": [10, 61], "sensor": [11], "signals": [12], "as": [13], "input.": [14], "In": [15], "this": [16, 39, 56], "study,": [17], "we": [18, 42, 65], "built": [19], "pretrained": [21, 40], "language": [22, 35], "model": [23, 103, 114], "based": [24], "on": [25, 47], "the": [26, 48, 93, 98, 110], "Transformer": [27, 72], "architecture,": [28], "which": [29], "is": [30], "widely": [31], "used": [32], "in": [33, 92, 118], "natural": [34], "processing.": [36], "By": [37], "leveraging": [38], "model,": [41], "aimed": [43], "to": [44, 109], "improve": [45], "performance": [46], "downstream": [49], "task": [50, 57], "of": [51, 100], "recognition.": [54], "While": [55], "can": [58], "be": [59], "addressed": [60], "vanilla": [63, 111], "Transformer,": [64, 112], "propose": [66], "an": [67], "enhanced": [68], "n-dimensional": [69, 79], "numerical": [70, 80], "processing": [71], "that": [73], "incorporates": [74], "three": [75], "key": [76], "features:": [77], "embedding": [78], "data": [81], "through": [82], "linear": [84, 90], "layer,": [85], "binning-based": [86], "pre-processing,": [87], "and": [88], "transformation": [91], "output": [94], "layer.": [95], "evaluated": [97], "effectiveness": [99], "our": [101, 113], "proposed": [102], "across": [104], "five": [105], "different": [106], "datasets.": [107], "Compared": [108], "demonstrated": [115], "10%-15%": [116], "improvements": [117], "accuracy.": [119]}, "primary_topic": {"id": "https://openalex.org/T10444", "display_name": "Context-Aware Activity Recognition Systems", "score": 0.986299991607666, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10444", "display_name": "Context-Aware Activity Recognition Systems", "score": 0.986299991607666, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12205", "display_name": "Time Series Analysis and Forecasting", "score": 0.003599999938160181, "subfield": {"id": "https://openalex.org/subfields/1711", "display_name": "Signal Processing"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.001500000013038516, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2505.21918", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2505.21918", "pdf_url": "https://arxiv.org/pdf/2505.21918", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417069110", "doi": "https://doi.org/10.48550/arxiv.2503.23306", "title": "Focus Directions Make Your Language Models Pay More Attention to Relevant Contexts", "publication_year": 2025, "publication_date": "2025-03-30", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5010382904", "display_name": "Youxiang Zhu", "orcid": "https://orcid.org/0009-0000-7294-1596", "observed_orcids": ["https://orcid.org/0009-0000-7294-1596"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Youxiang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Li, Ruochen", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Ruochen", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5090003719", "display_name": "Danqing Wang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Danqing", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5062859783", "display_name": "Daniel Haehn", "orcid": "https://orcid.org/0000-0001-9144-3461", "observed_orcids": ["https://orcid.org/0000-0001-9144-3461"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Haehn, Daniel", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5101655447", "display_name": "Xiaohui Liang", "orcid": "https://orcid.org/0000-0003-4064-2393", "observed_orcids": ["https://orcid.org/0000-0003-4064-2393"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liang, Xiaohui", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Long-context": [0], "large": [1], "language": [2], "models": [3], "(LLMs)": [4], "are": [5], "prone": [6], "to": [7, 53, 57, 67, 88, 92, 121], "be": [8, 62], "distracted": [9], "by": [10, 64], "irrelevant": [11], "contexts.": [12, 69], "The": [13], "reason": [14], "for": [15], "distraction": [16, 47], "remains": [17], "poorly": [18], "understood.": [19], "In": [20], "this": [21], "paper,": [22], "we": [23, 44], "first": [24], "identify": [25, 72], "the": [26, 37, 41, 77, 105, 123, 128], "contextual": [27, 50], "heads,": [28, 84], "a": [29], "special": [30], "group": [31], "of": [32, 40, 82, 107, 127], "attention": [33, 39, 56, 66, 91], "heads": [34, 51], "that": [35, 46], "control": [36], "overall": [38], "LLM.": [42], "Then,": [43], "demonstrate": [45], "arises": [48], "when": [49], "fail": [52], "allocate": [54, 89], "sufficient": [55], "relevant": [58, 93], "contexts": [59, 94], "and": [60, 79, 114], "can": [61], "mitigated": [63], "increasing": [65], "these": [68, 83], "We": [70, 102, 131], "further": [71, 137], "focus": [73, 108, 117], "directions,": [74], "located": [75], "at": [76], "key": [78], "query": [80], "activations": [81], "which": [85, 98], "enable": [86], "them": [87], "more": [90], "without": [95], "explicitly": [96], "specifying": [97], "context": [99], "is": [100], "relevant.": [101], "comprehensively": [103], "evaluate": [104], "effect": [106], "direction": [109], "on": [110, 139], "various": [111], "long-context": [112, 129, 140], "tasks": [113], "find": [115], "out": [116], "directions": [118], "could": [119, 135], "help": [120], "mitigate": [122], "poor": [124], "task": [125], "alignment": [126], "LLMs.": [130], "believe": [132], "our": [133], "findings": [134], "promote": [136], "research": [138], "LLM": [141], "alignment.": [142]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9369999766349792, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9369999766349792, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.02410000003874302, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.021299999207258224, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.23306", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.23306", "pdf_url": "https://arxiv.org/pdf/2503.23306", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4409474254", "doi": "https://doi.org/10.1109/jssc.2025.3554699", "title": "C-Transformer: An Energy-Efficient Homogeneous DNN-Transformer/SNN-Transformer Processor for Large Language Models", "publication_year": 2025, "publication_date": "2025-04-15", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100748588", "display_name": "Sangyeob Kim", "orcid": "https://orcid.org/0000-0002-1783-5296", "observed_orcids": ["https://orcid.org/0000-0002-1783-5296", "https://orcid.org/0000-0001-7639-0664"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Sangyeob Kim", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0000-0002-1783-5296", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100748581", "display_name": "Sangjin Kim", "orcid": "https://orcid.org/0000-0001-6665-9973", "observed_orcids": ["https://orcid.org/0000-0001-6665-9973"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Sangjin Kim", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0000-0001-6665-9973", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015963917", "display_name": "Wooyoung Jo", "orcid": "https://orcid.org/0000-0003-3598-3572", "observed_orcids": ["https://orcid.org/0000-0003-3598-3572"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Wooyoung Jo", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0000-0003-3598-3572", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100350562", "display_name": "Soyeon Kim", "orcid": "https://orcid.org/0000-0001-7049-9191", "observed_orcids": ["https://orcid.org/0000-0001-7049-9191"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Soyeon Kim", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0000-0001-7049-9191", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5086945425", "display_name": "Seongyon Hong", "orcid": "https://orcid.org/0009-0002-8891-735X", "observed_orcids": ["https://orcid.org/0009-0002-8891-735X"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Seongyon Hong", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0009-0002-8891-735X", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5019672143", "display_name": "Nayeong Lee", "orcid": "https://orcid.org/0009-0003-5439-2206", "observed_orcids": ["https://orcid.org/0009-0003-5439-2206"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Nayeong Lee", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0009-0003-5439-2206", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045558338", "display_name": "Jungwan Lee", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Jungwan Lee", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5077896259", "display_name": "Hoi\u2010Jun Yoo", "orcid": "https://orcid.org/0000-0002-6661-4879", "observed_orcids": ["https://orcid.org/0000-0002-6661-4879", "https://orcid.org/0000-0001-9323-0398"]}, "institutions": [{"id": "https://openalex.org/I157485424", "display_name": "Korea Advanced Institute of Science and Technology", "ror": "https://ror.org/05apxxy63", "country_code": "KR", "type": "education", "lineage": ["https://openalex.org/I157485424"]}], "countries": ["KR"], "is_corresponding": false, "raw_author_name": "Hoi-Jun Yoo", "raw_affiliation_strings": ["School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea"], "raw_orcid": "https://orcid.org/0000-0002-6661-4879", "affiliations": [{"raw_affiliation_string": "School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea", "institution_ids": ["https://openalex.org/I157485424"]}]}], "abstract_inverted_index": {"In": [0], "this": [1], "article,": [2], "we": [3], "propose": [4], "a": [5, 36, 90, 97, 175], "new": [6], "language": [7, 20, 109, 184, 219], "model": [8, 185], "processor": [9], "named": [10], "the": [11, 15, 53, 73, 135, 150, 160, 207], "C-Transformer": [12, 136, 180, 208], "to": [13, 41, 51, 71, 149, 224], "address": [14], "external": [16, 74, 143, 214], "memory": [17, 75, 144, 215], "bottleneck": [18], "of": [19, 24, 56, 93, 100], "models.": [21], "It": [22], "consists": [23], "three": [25], "key": [26], "functional": [27], "blocks:": [28], "1)": [29], "homogeneous": [30], "deep-neural-network": [31], "(DNN)\u2013Transformer/spiking-Transformer": [32], "core": [33], "(HDSC)": [34], "with": [35, 66, 96, 174, 187], "hybrid": [37], "multiplication/accumulation": [38], "unit": [39, 49, 64], "(HMAU)": [40], "enhance": [42], "hardware": [43], "utilization;": [44], "2)": [45], "output": [46], "spike": [47], "speculation": [48], "(OSSU)": [50], "increase": [52], "energy": [54, 140, 147, 157, 173, 211], "efficiency": [55], "spike-domain": [57], "processing;": [58], "and": [59, 87, 103, 112, 131, 141, 196, 204, 217], "3)": [60], "implicit": [61], "weight": [62], "generation": [63], "(IWGU)": [65], "extended": [67], "sign": [68], "compression": [69], "(ESC)": [70], "eliminate": [72], "bottleneck.": [76], "The": [77, 179], "chip": [78, 153], "is": [79], "fabricated": [80], "in": [81, 193, 198], "Samsung\u2019s": [82], "28-nm": [83], "1P8M": [84], "CMOS": [85], "technology": [86], "operates": [88], "at": [89], "supply": [91], "voltage": [92], "0.7\u20131.1": [94], "V": [95], "maximum": [98], "frequency": [99], "200": [101], "MHz": [102], "supports": [104], "various": [105, 183], "tasks": [106, 186], "such": [107, 221], "as": [108, 222], "modeling,": [110], "translation,": [111], "summarization.": [113], "For": [114], "models": [115, 220], "like": [116], "generative": [117], "pre-trained": [118], "transformer": [119, 125, 129], "2": [120], "(GPT-2),": [121], "multilingual": [122], "text-to-text": [123, 127], "transfer": [124, 128], "(mT5),": [126], "(T5),": [130], "FairSeq": [132], "MachineTranslation": [133], "(FSMT),": [134], "achieves": [137], "0.21\u00d7\u20130.33\u00d7": [138], "computation": [139], "0.37\u00d7\u20130.41\u00d7": [142], "access": [145], "(EMA)": [146], "compared": [148], "baseline.": [151], "Our": [152], "demonstrates": [154], "13.6%": [155], "lower": [156], "consumption": [158], "than": [159], "previous": [161], "state-of-the-art,": [162], "despite": [163], "having": [164], "2.1\u00d7": [165], "more": [166], "parameters.": [167], "Moreover,": [168], "it": [169], "consumes": [170], "63.8%": [171], "less": [172], "similar": [176], "parameter": [177], "size.": [178], "can": [181], "complete": [182], "<1": [188], "s": [189, 195], "latency,": [190], "notably": [191], "FSMT": [192], "0.09": [194], "GPT-2": [197, 223], "0.656": [199], "s.": [200], "By": [201], "combining": [202], "DNN\u2013Transformer": [203], "Spiking-Transformer": [205], "architectures,": [206], "enhances": [209], "computational": [210], "efficiency,": [212], "eliminates": [213], "bottlenecks,": [216], "enables": [218], "achieve": [225], "state-of-the-art": [226], "performance": [227], "on": [228], "mobile": [229], "devices.": [230]}, "primary_topic": {"id": "https://openalex.org/T10502", "display_name": "Advanced Memory and Neural Computing", "score": 0.365200012922287, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10502", "display_name": "Advanced Memory and Neural Computing", "score": 0.365200012922287, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10363", "display_name": "Low-power high-performance VLSI design", "score": 0.2996000051498413, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.15209999680519104, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1109/jssc.2025.3554699", "is_oa": false, "landing_page_url": "https://doi.org/10.1109/jssc.2025.3554699", "pdf_url": null, "source": {"id": "https://openalex.org/S83637746", "display_name": "IEEE Journal of Solid-State Circuits", "issn_l": "0018-9200", "issn": ["0018-9200", "1558-173X"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-3", "norway-2"], "host_organization": "https://openalex.org/P4310319808", "host_organization_name": "Institute of Electrical and Electronics Engineers", "host_organization_lineage": ["https://openalex.org/P4310319808"], "host_organization_lineage_names": ["Institute of Electrical and Electronics Engineers"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "IEEE Journal of Solid-State Circuits", "raw_type": "journal-article"}, "cited_by_count": 11, "is_retracted": false}, {"id": "https://openalex.org/W4415312617", "doi": "https://doi.org/10.48550/arxiv.2506.14646", "title": "GuiLoMo: Allocating Expert Number and Rank for LoRA-MoE via Bilevel Optimization with GuidedSelection Vectors", "publication_year": 2025, "publication_date": "2025-06-17", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5101811914", "display_name": "Hengyuan Zhang", "orcid": "https://orcid.org/0000-0001-5965-201X", "observed_orcids": ["https://orcid.org/0000-0001-5965-201X"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhang, Hengyuan", "raw_affiliation_strings": ["Tsinghua University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University ,", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063203224", "display_name": "Xinrong Chen", "orcid": "https://orcid.org/0000-0002-4186-0678", "observed_orcids": ["https://orcid.org/0000-0002-4186-0678", "https://orcid.org/0009-0005-7493-947X"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chen, Xinrong", "raw_affiliation_strings": ["Peking University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Peking University ,", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Qiu, Yingmin", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I139759216", "display_name": "Beijing University of Posts and Telecommunications", "ror": "https://ror.org/04w9fbh59", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I139759216"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Qiu, Yingmin", "raw_affiliation_strings": ["Beijing University of Posts and Telecommunications ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing University of Posts and Telecommunications ,", "institution_ids": ["https://openalex.org/I139759216"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101870507", "display_name": "Xiao Liang", "orcid": "https://orcid.org/0000-0002-3064-6908", "observed_orcids": ["https://orcid.org/0000-0002-3064-6908"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Liang, Xiao", "raw_affiliation_strings": ["Tsinghua University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Tsinghua University ,", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111259477", "display_name": "Ziyue Li", "orcid": "https://orcid.org/0009-0004-3677-2886", "observed_orcids": ["https://orcid.org/0009-0004-3677-2886"]}, "institutions": [{"id": "https://openalex.org/I4210148850", "display_name": "National Science Library", "ror": "https://ror.org/04ndjyr10", "country_code": "CN", "type": "archive", "lineage": ["https://openalex.org/I19820366", "https://openalex.org/I4210148850"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Li, Ziyue", "raw_affiliation_strings": ["National Science Library , Chinese Academy of Sciences , 5 Baidu ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "National Science Library , Chinese Academy of Sciences , 5 Baidu ,", "institution_ids": ["https://openalex.org/I4210148850"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5102908620", "display_name": "Guanyu Wang", "orcid": "https://orcid.org/0009-0009-4048-9422", "observed_orcids": ["https://orcid.org/0009-0009-4048-9422"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Guanyu", "raw_affiliation_strings": ["Peking University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Peking University ,", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100415567", "display_name": "Wei-Ping Li", "orcid": "https://orcid.org/0000-0002-9171-6439", "observed_orcids": ["https://orcid.org/0000-0002-9171-6439"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Li, Weiping", "raw_affiliation_strings": ["Peking University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Peking University ,", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5059356240", "display_name": "Tong Mo", "orcid": "https://orcid.org/0000-0002-3564-4610", "observed_orcids": ["https://orcid.org/0000-0002-3564-4610"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Mo, Tong", "raw_affiliation_strings": ["Peking University ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Peking University ,", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020581824", "display_name": "Hayden Kwok\u2010Hay So", "orcid": "https://orcid.org/0000-0002-6514-0237", "observed_orcids": ["https://orcid.org/0000-0002-6514-0237"]}, "institutions": [{"id": "https://openalex.org/I889458895", "display_name": "University of Hong Kong", "ror": "https://ror.org/02zhqgq86", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I889458895"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "So, Hayden Kwok-Hay", "raw_affiliation_strings": ["The University of Hong Kong"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The University of Hong Kong", "institution_ids": ["https://openalex.org/I889458895"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5072120057", "display_name": "Ngai Wong", "orcid": "https://orcid.org/0000-0002-3026-0108", "observed_orcids": ["https://orcid.org/0000-0002-3026-0108"]}, "institutions": [{"id": "https://openalex.org/I889458895", "display_name": "University of Hong Kong", "ror": "https://ror.org/02zhqgq86", "country_code": "HK", "type": "education", "lineage": ["https://openalex.org/I889458895"]}], "countries": ["HK"], "is_corresponding": false, "raw_author_name": "Wong, Ngai", "raw_affiliation_strings": ["The University of Hong Kong"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "The University of Hong Kong", "institution_ids": ["https://openalex.org/I889458895"]}]}], "abstract_inverted_index": {"Parameter-efficient": [0], "fine-tuning": [1], "(PEFT)": [2], "methods,": [3], "particularly": [4], "Low-Rank": [5], "Adaptation": [6], "(LoRA),": [7], "offer": [8], "an": [9], "efficient": [10], "way": [11], "to": [12, 43, 111, 122, 146], "adapt": [13], "large": [14], "language": [15], "models": [16, 133], "with": [17, 37, 98], "reduced": [18], "computational": [19], "costs.": [20], "However,": [21], "their": [22], "performance": [23, 145], "is": [24, 174], "limited": [25], "by": [26], "the": [27, 38, 52, 59, 70, 166], "small": [28], "number": [29], "of": [30, 55, 61, 168], "trainable": [31], "parameters.": [32], "Recent": [33], "work": [34], "combines": [35], "LoRA": [36, 76], "Mixture-of-Experts": [39], "(MoE),": [40], "i.e.,": [41], "LoRA-MoE,": [42], "enhance": [44], "capacity,": [45], "but": [46], "two": [47], "limitations": [48], "remain": [49], "in": [50], "hindering": [51], "full": [53], "exploitation": [54], "its": [56], "potential:": [57], "1)": [58], "influence": [60], "downstream": [62], "tasks": [63], "when": [64], "assigning": [65], "expert": [66, 92, 125, 156, 170], "numbers,": [67], "and": [68, 94, 115, 118, 127, 158, 163], "2)": [69], "uniform": [71], "rank": [72], "assignment": [73], "across": [74, 134, 161], "all": [75, 147], "experts,": [77], "which": [78], "restricts": [79], "representational": [80], "diversity.": [81], "To": [82], "mitigate": [83], "these": [84], "gaps,": [85], "we": [86], "propose": [87], "GuiLoMo,": [88], "a": [89, 106], "fine-grained": [90], "layer-wise": [91], "numbers": [93, 126, 157], "ranks": [95, 159], "allocation": [96], "strategy": [97], "GuidedSelection": [99], "Vectors": [100], "(GSVs).": [101], "GSVs": [102], "are": [103, 119], "learned": [104], "via": [105], "prior": [107], "bilevel": [108], "optimization": [109], "process": [110], "capture": [112], "both": [113], "model-": [114], "task-specific": [116], "needs,": [117], "then": [120], "used": [121], "allocate": [123], "optimal": [124], "ranks.": [128], "Experiments": [129], "on": [130], "three": [131], "backbone": [132], "diverse": [135], "benchmarks": [136], "show": [137], "that": [138], "GuiLoMo": [139], "consistently": [140], "achieves": [141], "superior": [142], "or": [143], "comparable": [144], "baselines.": [148], "Further": [149], "analysis": [150], "offers": [151], "key": [152], "insights": [153], "into": [154], "how": [155], "vary": [160], "layers": [162], "tasks,": [164], "highlighting": [165], "benefits": [167], "adaptive": [169], "configuration.": [171], "Our": [172], "code": [173], "available": [175], "at": [176], "https://github.com/Liar406/Gui-LoMo.git.": [177]}, "primary_topic": {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.9535999894142151, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.9535999894142151, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.03290000185370445, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.0052999998442828655, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.14646", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.14646", "pdf_url": "https://arxiv.org/pdf/2506.14646", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415330535", "doi": "https://doi.org/10.48550/arxiv.2508.04664", "title": "Sculptor: Empowering LLMs with Cognitive Agency via Active Context Management", "publication_year": 2025, "publication_date": "2025-08-06", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Li, Mo", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Mo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5072024833", "display_name": "Lihua Xu", "orcid": "https://orcid.org/0000-0002-2237-9336", "observed_orcids": ["https://orcid.org/0000-0002-2237-9336", "https://orcid.org/0000-0002-2272-2931"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Xu, L. H.", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Tan, Qitai", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tan, Qitai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5113190251", "display_name": "Long Ma", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Ma, Long", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070897302", "display_name": "Ting Cao", "orcid": "https://orcid.org/0000-0002-2241-7742", "observed_orcids": ["https://orcid.org/0000-0002-2241-7742"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cao, Ting", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5027782298", "display_name": "Yunxin Liu", "orcid": "https://orcid.org/0000-0003-1333-4619", "observed_orcids": ["https://orcid.org/0000-0003-1333-4619", "https://orcid.org/0000-0001-7352-8955"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Yunxin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs)": [3], "suffer": [4], "from": [5], "significant": [6], "performance": [7, 124], "degradation": [8], "when": [9], "processing": [10], "long": [11], "contexts": [12], "due": [13], "to": [14, 39, 56, 92, 101, 201], "proactive": [15, 173], "interference,": [16], "where": [17], "irrelevant": [18], "information": [19, 108], "in": [20], "earlier": [21], "parts": [22], "of": [23, 74, 154], "the": [24, 152], "context": [25, 77], "disrupts": [26], "reasoning": [27, 184], "and": [28, 82, 84, 97, 133], "memory": [29, 37], "recall.": [30], "While": [31], "most": [32], "research": [33], "focuses": [34], "on": [35, 106, 115], "external": [36], "systems": [38], "augment": [40], "LLMs'": [41, 130], "capabilities,": [42], "we": [43, 141], "propose": [44], "a": [45, 66, 143, 178], "complementary": [46], "approach:": [47], "empowering": [48], "LLMs": [49, 70, 91], "with": [50, 71], "Active": [51, 166], "Context": [52, 167], "Management": [53], "(ACM)": [54], "tools": [55], "actively": [57, 158], "sculpt": [58], "their": [59, 95], "internal": [60], "working": [61, 98], "memory.": [62], "We": [63], "introduce": [64, 142], "Sculptor,": [65], "framework": [67], "that": [68, 120, 157, 189], "equips": [69], "three": [72], "categories": [73], "tools:": [75], "(1)": [76], "fragmentation,": [78], "(2)": [79], "summary,": [80], "hide,": [81], "restore,": [83], "(3)": [85], "precise": [86], "search.": [87], "Our": [88], "approach": [89], "enables": [90], "proactively": [93], "manage": [94], "attention": [96], "memory,": [99], "analogous": [100], "how": [102], "humans": [103], "selectively": [104], "focus": [105], "relevant": [107], "while": [109], "filtering": [110], "out": [111], "distractions.": [112], "Experimental": [113], "evaluation": [114], "diverse": [116, 186], "long-context": [117, 187], "benchmarks": [118], "demonstrates": [119], "Sculptor": [121, 169], "significantly": [122], "improves": [123], "even": [125], "without": [126], "specific": [127], "training,": [128], "leveraging": [129], "inherent": [131], "tool-calling": [132], "instruction-following": [134], "capabilities.": [135], "To": [136], "further": [137], "optimize": [138], "these": [139], "strategies,": [140, 192], "novel": [144], "dynamic": [145], "context-aware": [146], "reinforcement": [147], "learning": [148], "(RL)": [149], "approach,": [150], "advancing": [151], "training": [153], "an": [155], "agent": [156], "modifies": [159], "its": [160], "own": [161], "conversational": [162], "history.": [163], "By": [164], "enabling": [165], "Management,": [168], "not": [170], "only": [171], "mitigates": [172], "interference": [174], "but": [175], "also": [176], "provides": [177], "cognitive": [179], "foundation": [180], "for": [181], "more": [182], "reliable": [183], "across": [185], "tasks-highlighting": [188], "explicit": [190], "context-control": [191], "rather": [193], "than": [194], "merely": [195], "larger": [196], "token": [197], "windows,": [198], "are": [199], "key": [200], "robustness": [202], "at": [203], "scale.": [204]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7372999787330627, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.7372999787330627, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.09390000253915787, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.06199999898672104, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2508.04664", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2508.04664", "pdf_url": "https://arxiv.org/pdf/2508.04664", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414151467", "doi": "https://doi.org/10.2139/ssrn.5409611", "title": "How LLMs Evolved into Emergence", "publication_year": 2025, "publication_date": "2025-01-01", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5087458971", "display_name": "Fred Amonya", "orcid": "https://orcid.org/0000-0002-4859-0113", "observed_orcids": ["https://orcid.org/0000-0002-4859-0113"]}, "institutions": [{"id": "https://openalex.org/I79619799", "display_name": "University of Birmingham", "ror": "https://ror.org/03angcq70", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I79619799"]}], "countries": ["GB"], "is_corresponding": true, "raw_author_name": "Fred Amonya", "raw_affiliation_strings": ["University of Birmingham"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Birmingham", "institution_ids": ["https://openalex.org/I79619799"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T13308", "display_name": "Cybernetics and Technology in Society", "score": 0.8833000063896179, "subfield": {"id": "https://openalex.org/subfields/1207", "display_name": "History and Philosophy of Science"}, "field": {"id": "https://openalex.org/fields/12", "display_name": "Arts and Humanities"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T13308", "display_name": "Cybernetics and Technology in Society", "score": 0.8833000063896179, "subfield": {"id": "https://openalex.org/subfields/1207", "display_name": "History and Philosophy of Science"}, "field": {"id": "https://openalex.org/fields/12", "display_name": "Arts and Humanities"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.04340000078082085, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12797", "display_name": "History of Computing Technologies", "score": 0.011900000274181366, "subfield": {"id": "https://openalex.org/subfields/1706", "display_name": "Computer Science Applications"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.2139/ssrn.5409611", "is_oa": true, "landing_page_url": "https://doi.org/10.2139/ssrn.5409611", "pdf_url": null, "source": {"id": "https://openalex.org/S4210172589", "display_name": "SSRN Electronic Journal", "issn_l": "1556-5068", "issn": ["1556-5068"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I1318003438", "host_organization_name": "RELX Group (Netherlands)", "host_organization_lineage": ["https://openalex.org/I1318003438"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415616656", "doi": "https://doi.org/10.1504/ijcat.2025.149370", "title": "Large model driven adaptive English translation mechanism", "publication_year": 2025, "publication_date": "2025-01-01", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5075511851", "display_name": "Haiying Song", "orcid": "https://orcid.org/0000-0001-7454-0542", "observed_orcids": ["https://orcid.org/0000-0001-7454-0542"]}, "institutions": [{"id": "https://openalex.org/I92092256", "display_name": "Huanghe Science and Technology College", "ror": "https://ror.org/008p6rr25", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I92092256"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Haiying Song", "raw_affiliation_strings": ["Huanghe University of Science and Technology, Zhengzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Huanghe University of Science and Technology, Zhengzhou, China", "institution_ids": ["https://openalex.org/I92092256"]}]}], "abstract_inverted_index": {"Significant": [0], "advancements": [1], "have": [2], "been": [3], "made": [4], "in": [5, 27, 79, 96, 120], "the": [6, 14, 109, 115], "field": [7], "of": [8, 16, 111, 117], "natural": [9], "language": [10, 18, 35], "processing": [11], "thanks": [12], "to": [13, 68, 103], "development": [15], "large-scale": [17], "models.": [19, 105], "However,": [20], "existing": [21], "translation": [22, 46, 123], "methods": [23], "often": [24], "face": [25], "limitations": [26], "handling": [28], "contextual": [29], "nuances,": [30], "domain-specific": [31, 56], "terminology": [32], "and": [33, 58, 77, 91, 99, 134], "evolving": [34], "usage.": [36], "To": [37], "address": [38], "these": [39], "challenges,": [40], "we": [41], "propose": [42], "an": [43], "adaptive": [44, 53, 118, 135], "English": [45], "mechanism": [47], "based": [48], "on": [49, 84], "large": [50], "models,": [51], "incorporating": [52, 73], "feedback": [54, 76], "integration,": [55], "fine-tuning": [57], "a": [59], "dynamic": [60], "learning": [61], "mechanism.": [62], "Our": [63], "approach": [64], "leverages": [65], "Transformer": [66], "architectures": [67], "continuously": [69], "refine": [70], "translations": [71], "by": [72], "real-time": [74], "user": [75], "specialising": [78], "various": [80], "domains.": [81], "Evaluations": [82], "conducted": [83], "data": [85], "sets": [86], "including": [87], "WMT,": [88], "OpenSubtitles,": [89], "IWSLT": [90], "CCMT": [92], "demonstrate": [93], "significant": [94], "improvements": [95], "BLEU,": [97], "METEOR": [98], "TER": [100], "scores": [101], "compared": [102], "other": [104], "Ablation": [106], "studies": [107], "confirm": [108], "contributions": [110], "each": [112], "component,": [113], "highlighting": [114], "importance": [116], "mechanisms": [119], "achieving": [121], "superior": [122], "performance.": [124], "These": [125], "findings": [126], "indicate": [127], "that": [128], "our": [129], "proposed": [130], "method": [131], "offers": [132], "robust": [133], "solutions": [136], "for": [137], "diverse": [138], "linguistic": [139], "applications.": [140]}, "primary_topic": {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.9868000149726868, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.009399999864399433, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10201", "display_name": "Speech Recognition and Synthesis", "score": 0.0017000000225380063, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1504/ijcat.2025.149370", "is_oa": false, "landing_page_url": "https://doi.org/10.1504/ijcat.2025.149370", "pdf_url": null, "source": {"id": "https://openalex.org/S4210207686", "display_name": "International Journal of Computer Applications in Technology", "issn_l": "0952-8091", "issn": ["0952-8091", "1741-5047"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310317825", "host_organization_name": "Inderscience Publishers", "host_organization_lineage": ["https://openalex.org/P4310317825"], "host_organization_lineage_names": ["Inderscience Publishers"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "International Journal of Computer Applications in Technology", "raw_type": "journal-article"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415208802", "doi": "https://doi.org/10.1038/s42256-025-01126-3", "title": "A neural symbolic model for space physics", "publication_year": 2025, "publication_date": "2025-10-15", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5089391283", "display_name": "Jie Ying", "orcid": "https://orcid.org/0009-0003-7044-2242", "observed_orcids": ["https://orcid.org/0009-0003-7044-2242"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Jie Ying", "raw_affiliation_strings": ["Qiuzhen College, Tsinghua University, Beijing, China"], "raw_orcid": "https://orcid.org/0009-0003-7044-2242", "affiliations": [{"raw_affiliation_string": "Qiuzhen College, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101191761", "display_name": "Haowei Lin", "orcid": "https://orcid.org/0009-0006-9809-4835", "observed_orcids": ["https://orcid.org/0009-0006-9809-4835"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Haowei Lin", "raw_affiliation_strings": ["Institute for Artificial Intelligence, Peking University, Beijing, China"], "raw_orcid": "https://orcid.org/0009-0006-9809-4835", "affiliations": [{"raw_affiliation_string": "Institute for Artificial Intelligence, Peking University, Beijing, China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5031323973", "display_name": "Chao Yue", "orcid": "https://orcid.org/0000-0001-9720-5210", "observed_orcids": ["https://orcid.org/0000-0001-9720-5210"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chao Yue", "raw_affiliation_strings": ["School of Earth and Space Sciences, Peking University, Beijing, China"], "raw_orcid": "https://orcid.org/0000-0001-9720-5210", "affiliations": [{"raw_affiliation_string": "School of Earth and Space Sciences, Peking University, Beijing, China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5015501344", "display_name": "Yajie Chen", "orcid": "https://orcid.org/0000-0001-5494-4339", "observed_orcids": ["https://orcid.org/0000-0001-5494-4339"]}, "institutions": [{"id": "https://openalex.org/I4210116419", "display_name": "Max Planck Institute for Solar System Research", "ror": "https://ror.org/02j6gm739", "country_code": "DE", "type": "facility", "lineage": ["https://openalex.org/I149899117", "https://openalex.org/I4210116419"]}], "countries": ["DE"], "is_corresponding": false, "raw_author_name": "Yajie Chen", "raw_affiliation_strings": ["Max Planck Institute for Solar System Research, G\u00f6ttingen, Germany"], "raw_orcid": "https://orcid.org/0000-0001-5494-4339", "affiliations": [{"raw_affiliation_string": "Max Planck Institute for Solar System Research, G\u00f6ttingen, Germany", "institution_ids": ["https://openalex.org/I4210116419"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5038718275", "display_name": "Chao Xiao", "orcid": "https://orcid.org/0000-0002-2633-9526", "observed_orcids": ["https://orcid.org/0000-0002-2633-9526"]}, "institutions": [{"id": "https://openalex.org/I154099455", "display_name": "Shandong University", "ror": "https://ror.org/0207yh398", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I154099455"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Chao Xiao", "raw_affiliation_strings": ["Institute of Space Sciences, Shandong University, Weihai, China"], "raw_orcid": "https://orcid.org/0000-0002-2633-9526", "affiliations": [{"raw_affiliation_string": "Institute of Space Sciences, Shandong University, Weihai, China", "institution_ids": ["https://openalex.org/I154099455"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5080294388", "display_name": "Quanqi Shi", "orcid": "https://orcid.org/0000-0001-6835-4751", "observed_orcids": ["https://orcid.org/0000-0001-6835-4751"]}, "institutions": [{"id": "https://openalex.org/I154099455", "display_name": "Shandong University", "ror": "https://ror.org/0207yh398", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I154099455"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Quanqi Shi", "raw_affiliation_strings": ["Institute of Space Sciences, Shandong University, Weihai, China"], "raw_orcid": "https://orcid.org/0000-0001-6835-4751", "affiliations": [{"raw_affiliation_string": "Institute of Space Sciences, Shandong University, Weihai, China", "institution_ids": ["https://openalex.org/I154099455"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5112110952", "display_name": "Yitao Liang", "orcid": "https://orcid.org/0009-0009-0944-7814", "observed_orcids": ["https://orcid.org/0009-0009-0944-7814"]}, "institutions": [{"id": "https://openalex.org/I20231570", "display_name": "Peking University", "ror": "https://ror.org/02v51f717", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I20231570"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yitao Liang", "raw_affiliation_strings": ["Institute for Artificial Intelligence, Peking University, Beijing, China"], "raw_orcid": "https://orcid.org/0009-0009-0944-7814", "affiliations": [{"raw_affiliation_string": "Institute for Artificial Intelligence, Peking University, Beijing, China", "institution_ids": ["https://openalex.org/I20231570"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5006865040", "display_name": "Shing\u2013Tung Yau", "orcid": "https://orcid.org/0000-0003-3394-2187", "observed_orcids": ["https://orcid.org/0000-0003-3394-2187"]}, "institutions": [{"id": "https://openalex.org/I4403928416", "display_name": "Beijing Institute of Mathematical Sciences and Applications", "ror": "https://ror.org/05t6hvr95", "country_code": null, "type": "education", "lineage": ["https://openalex.org/I4403928416", "https://openalex.org/I99065089"]}, {"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Shing-Tung Yau", "raw_affiliation_strings": ["Beijing Institute of Mathematical Sciences and Applications, Beijing, China", "Yau Mathematical Sciences Center, Tsinghua University, Beijing, China"], "raw_orcid": "https://orcid.org/0000-0003-3394-2187", "affiliations": [{"raw_affiliation_string": "Beijing Institute of Mathematical Sciences and Applications, Beijing, China", "institution_ids": ["https://openalex.org/I4403928416"]}, {"raw_affiliation_string": "Yau Mathematical Sciences Center, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045927034", "display_name": "Yuan Zhou", "orcid": "https://orcid.org/0009-0008-1706-6539", "observed_orcids": ["https://orcid.org/0009-0008-1706-6539"]}, "institutions": [{"id": "https://openalex.org/I4403928416", "display_name": "Beijing Institute of Mathematical Sciences and Applications", "ror": "https://ror.org/05t6hvr95", "country_code": null, "type": "education", "lineage": ["https://openalex.org/I4403928416", "https://openalex.org/I99065089"]}, {"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Yuan Zhou", "raw_affiliation_strings": ["Beijing Institute of Mathematical Sciences and Applications, Beijing, China", "Department of Mathematical Sciences, Tsinghua University, Beijing, China", "Yau Mathematical Sciences Center, Tsinghua University, Beijing, China"], "raw_orcid": "https://orcid.org/0009-0008-1706-6539", "affiliations": [{"raw_affiliation_string": "Beijing Institute of Mathematical Sciences and Applications, Beijing, China", "institution_ids": ["https://openalex.org/I4403928416"]}, {"raw_affiliation_string": "Department of Mathematical Sciences, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}, {"raw_affiliation_string": "Yau Mathematical Sciences Center, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5074269040", "display_name": "Jianzhu Ma", "orcid": "https://orcid.org/0000-0002-8236-6609", "observed_orcids": ["https://orcid.org/0000-0002-8236-6609"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Jianzhu Ma", "raw_affiliation_strings": ["Department of Electronic Engineering, Tsinghua University, Beijing, China", "Institute for AI Industry Research, Tsinghua University, Beijing, China"], "raw_orcid": "https://orcid.org/0000-0002-8236-6609", "affiliations": [{"raw_affiliation_string": "Department of Electronic Engineering, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}, {"raw_affiliation_string": "Institute for AI Industry Research, Tsinghua University, Beijing, China", "institution_ids": ["https://openalex.org/I99065089"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T11206", "display_name": "Model Reduction and Neural Networks", "score": 0.6137999892234802, "subfield": {"id": "https://openalex.org/subfields/3109", "display_name": "Statistical and Nonlinear Physics"}, "field": {"id": "https://openalex.org/fields/31", "display_name": "Physics and Astronomy"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11206", "display_name": "Model Reduction and Neural Networks", "score": 0.6137999892234802, "subfield": {"id": "https://openalex.org/subfields/3109", "display_name": "Statistical and Nonlinear Physics"}, "field": {"id": "https://openalex.org/fields/31", "display_name": "Physics and Astronomy"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11975", "display_name": "Evolutionary Algorithms and Applications", "score": 0.211899995803833, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11948", "display_name": "Machine Learning in Materials Science", "score": 0.019300000742077827, "subfield": {"id": "https://openalex.org/subfields/2505", "display_name": "Materials Chemistry"}, "field": {"id": "https://openalex.org/fields/25", "display_name": "Materials Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1038/s42256-025-01126-3", "is_oa": false, "landing_page_url": "https://doi.org/10.1038/s42256-025-01126-3", "pdf_url": null, "source": {"id": "https://openalex.org/S2912241403", "display_name": "Nature Machine Intelligence", "issn_l": "2522-5839", "issn": ["2522-5839"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-2", "ki-jl-3", "norway-2"], "host_organization": "https://openalex.org/P4310319908", "host_organization_name": "Nature Portfolio", "host_organization_lineage": ["https://openalex.org/P4310319908", "https://openalex.org/P4310319965"], "host_organization_lineage_names": ["Nature Portfolio", "Springer Nature"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Nature Machine Intelligence", "raw_type": "journal-article"}, "cited_by_count": 4, "is_retracted": false}, {"id": "https://openalex.org/W4416404598", "doi": "https://doi.org/10.1007/s12273-025-1324-9", "title": "Pre-training large language models based on Transformer architecture for building industry application: A review", "publication_year": 2025, "publication_date": "2025-11-01", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Jinrui Liang", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I37987034", "display_name": "Guangzhou University", "ror": "https://ror.org/05ar8rn06", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I37987034"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Jinrui Liang", "raw_affiliation_strings": ["School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "institution_ids": ["https://openalex.org/I37987034"]}, {"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China", "institution_ids": ["https://openalex.org/I37987034"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110703462", "display_name": "Yiquan Lin", "orcid": "https://orcid.org/0009-0006-1087-6016", "observed_orcids": ["https://orcid.org/0009-0006-1087-6016"]}, "institutions": [{"id": "https://openalex.org/I37987034", "display_name": "Guangzhou University", "ror": "https://ror.org/05ar8rn06", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I37987034"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yuyan Lin", "raw_affiliation_strings": ["School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "institution_ids": ["https://openalex.org/I37987034"]}, {"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China", "institution_ids": ["https://openalex.org/I37987034"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100677989", "display_name": "Yanchen Liu", "orcid": "https://orcid.org/0000-0001-9431-6919", "observed_orcids": ["https://orcid.org/0000-0001-9431-6919", "https://orcid.org/0009-0005-1258-1820"]}, "institutions": [{"id": "https://openalex.org/I37987034", "display_name": "Guangzhou University", "ror": "https://ror.org/05ar8rn06", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I37987034"]}], "countries": ["CN"], "is_corresponding": true, "raw_author_name": "Yanchen Liu", "raw_affiliation_strings": ["School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "institution_ids": ["https://openalex.org/I37987034"]}, {"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China", "institution_ids": ["https://openalex.org/I37987034"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070956286", "display_name": "Yu Huang", "orcid": "https://orcid.org/0000-0002-1138-1828", "observed_orcids": ["https://orcid.org/0000-0002-1138-1828"]}, "institutions": [{"id": "https://openalex.org/I37987034", "display_name": "Guangzhou University", "ror": "https://ror.org/05ar8rn06", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I37987034"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yu Huang", "raw_affiliation_strings": ["School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "institution_ids": ["https://openalex.org/I37987034"]}, {"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China", "institution_ids": ["https://openalex.org/I37987034"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5064346308", "display_name": "Huijun Wu", "orcid": "https://orcid.org/0000-0003-0322-9284", "observed_orcids": ["https://orcid.org/0000-0003-0322-9284"]}, "institutions": [{"id": "https://openalex.org/I37987034", "display_name": "Guangzhou University", "ror": "https://ror.org/05ar8rn06", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I37987034"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Huijun Wu", "raw_affiliation_strings": ["School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, 510006, China", "institution_ids": ["https://openalex.org/I37987034"]}, {"raw_affiliation_string": "School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, China", "institution_ids": ["https://openalex.org/I37987034"]}]}], "abstract_inverted_index": null, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.6029999852180481, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.6029999852180481, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11006", "display_name": "BIM and Construction Integration", "score": 0.3441999852657318, "subfield": {"id": "https://openalex.org/subfields/2215", "display_name": "Building and Construction"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T13904", "display_name": "Artificial Intelligence Applications", "score": 0.014100000262260437, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1007/s12273-025-1324-9", "is_oa": false, "landing_page_url": "https://doi.org/10.1007/s12273-025-1324-9", "pdf_url": null, "source": {"id": "https://openalex.org/S153850590", "display_name": "Building Simulation", "issn_l": "1996-3599", "issn": ["1996-3599", "1996-8744"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310319900", "host_organization_name": "Springer Science+Business Media", "host_organization_lineage": ["https://openalex.org/P4310319900", "https://openalex.org/P4310319965"], "host_organization_lineage_names": ["Springer Science+Business Media", "Springer Nature"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Building Simulation", "raw_type": "journal-article"}, "cited_by_count": 4, "is_retracted": false}, {"id": "https://openalex.org/W7124948954", "doi": "https://doi.org/10.3929/ethz-c-000791798", "title": "Pre-Attention Expert Prediction and Prefetching for Mixture-of-Experts Large Language Models", "publication_year": 2025, "publication_date": "2025-11-10", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5087479741", "display_name": "Shien Zhu", "orcid": "https://orcid.org/0000-0002-2094-7643", "observed_orcids": ["https://orcid.org/0000-0002-2094-7643"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Shien", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Bohl, Samuel", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bohl, Samuel", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5120361850", "display_name": "Robin Oester", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Oester, Robin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5123394977", "display_name": "Gustavo Alonso", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Alonso, Gustavo", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Mixture-of-Experts": [0], "(MoE)": [1], "Large": [2], "Language": [3], "Models": [4], "(LLMs)": [5], "efficiently": [6], "scale-up": [7], "the": [8, 23, 36, 45, 48, 58, 106, 120, 123, 127, 146, 176], "model": [9], "while": [10], "keeping": [11], "relatively": [12], "low": [13, 54], "inference": [14], "cost.": [15], "As": [16], "MoE": [17], "models": [18], "only": [19], "activate": [20], "part": [21], "of": [22, 108], "experts,": [24], "related": [25], "work": [26], "has": [27], "proposed": [28], "expert": [29, 83, 90, 152], "prediction": [30, 72, 84], "and": [31, 56, 88, 134, 164], "caching": [32], "methods": [33], "to": [34, 85, 137], "prefetch": [35], "experts": [37, 110], "for": [38, 51, 70], "faster": [39], "inference.": [40], "However,": [41], "existing": [42], "approaches": [43], "utilize": [44, 119], "activations": [46, 121], "from": [47], "previous": [49], "layer": [50, 60, 129], "prediction,": [52, 140], "incurring": [53], "accuracy": [55, 156, 174], "leave": [57], "first": [59, 147], "unoptimized.": [61], "Applying": [62], "complex": [63], "layers": [64], "or": [65], "even": [66], "training": [67], "standalone": [68], "networks": [69], "better": [71], "introduces": [73], "high": [74], "computation": [75], "overhead.": [76], "In": [77], "this": [78], "paper,": [79], "we": [80, 118], "propose": [81], "pre-attention": [82, 151], "achieve": [86, 138, 154], "accurate": [87, 139], "lightweight": [89], "prefetching.": [91], "The": [92], "key": [93], "insight": [94], "is": [95, 115], "that": [96, 104], "some": [97], "functions": [98, 114, 133], "in": [99, 126, 145], "LLMs": [100], "are": [101], "ranking-preserving,": [102], "indicating": [103], "matching": [105], "ranking": [107], "selected": [109], "using": [111], "simple": [112], "linear": [113, 132], "possible.": [116], "Therefore,": [117], "before": [122], "attention": [124], "block": [125], "same": [128], "with": [130], "2": [131], "ranking-aware": [135], "loss": [136], "which": [141], "also": [142], "supports": [143], "prefetching": [144], "layer.": [148], "Our": [149], "lightweight,": [150], "routers": [153], "93.03%": [155], "on": [157, 162, 166, 172], "DeepSeek": [158], "V2": [159], "Lite,": [160], "94.69%": [161], "Qwen3-30B,": [163], "97.62%": [165], "Phi-mini-MoE,": [167], "showing": [168], "about": [169], "15%": [170], "improvement": [171], "absolute": [173], "over": [175], "state-of-the-art": [177], "methods.": [178]}, "primary_topic": {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.6845999956130981, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.6845999956130981, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.1225999966263771, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.06840000301599503, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:www.research-collection.ethz.ch:20.500.11850/791798", "is_oa": true, "landing_page_url": "http://hdl.handle.net/20.500.11850/791798", "pdf_url": null, "source": {"id": "https://openalex.org/S4306402302", "display_name": "Repository for Publications and Research Data (ETH Zurich)", "issn_l": null, "issn": null, "is_oa": false, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I35440088", "host_organization_name": "ETH Zurich", "host_organization_lineage": ["https://openalex.org/I35440088"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by-nc-sa", "license_id": "https://openalex.org/licenses/cc-by-nc-sa", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": "arXiv", "raw_type": "info:eu-repo/semantics/workingPaper"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415067817", "doi": "https://doi.org/10.48550/arxiv.2506.11006", "title": "Test code generation at Ericsson using Program Analysis Augmented Fine Tuned LLMs", "publication_year": 2025, "publication_date": "2025-04-23", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5005909292", "display_name": "Sai Krishna", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Krishna, Sai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5025743698", "display_name": "Balvinder Singh", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Singh, Balvinder", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5064379712", "display_name": "Sujoy Roychowdhury", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Roychowdhury, Sujoy", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5072230453", "display_name": "Giriprasad Sridhara", "orcid": "https://orcid.org/0009-0003-7670-5701", "observed_orcids": ["https://orcid.org/0009-0003-7670-5701"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sridhara, Giriprasad", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5085520581", "display_name": "Sourav Mazumdar", "orcid": "https://orcid.org/0009-0006-6644-3914", "observed_orcids": ["https://orcid.org/0009-0006-6644-3914"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Mazumdar, Sourav", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119950100", "display_name": "Magnus Sandelin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sandelin, Magnus", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5042885262", "display_name": "Dimitris Rentas", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Rentas, Dimitris", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119950102", "display_name": "Maciej Nalepa", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Nalepa, Maciej", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119950103", "display_name": "Karol Sawicki", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sawicki, Karol", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5029162178", "display_name": "J. Gajda", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gajda, Jakub", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"We": [0, 33, 58, 91], "describe": [1, 34, 93], "test": [2, 16, 31, 150], "code": [3, 26, 56, 151], "generation": [4], "using": [5, 87], "Large": [6], "Language": [7], "Models": [8], "(LLMs)": [9], "in": [10, 18, 44, 54, 165], "Ericsson.": [11], "Our": [12, 131], "input": [13], "is": [14, 25, 107, 191], "a": [15, 67, 111, 172, 197], "step": [17], "natural": [19], "language": [20], "(English)": [21], "and": [22, 42, 48, 190], "our": [23, 135], "output": [24], "(Java)": [27], "which": [28, 50, 116], "accomplishes": [29], "the": [30, 55, 64, 80, 101, 141, 146, 155, 162, 166, 187, 194], "step.": [32], "how": [35, 61], "straight": [36], "forward": [37], "prompting": [38], "does": [39], "not": [40, 52], "suffice": [41], "results": [43, 132], "LLM": [45], "assuming": [46], "functions": [47], "signatures": [49], "are": [51], "present": [53], "repository.": [57], "then": [59, 92], "show": [60], "we": [62, 97], "alleviate": [63], "problem": [65], "by": [66, 99, 154], "combination": [68], "of": [69, 158, 171, 175, 184], "Retrieval": [70], "Augmented": [71], "Generation": [72], "(RAG)": [73], "along": [74], "with": [75, 83, 145], "prompt": [76, 82, 114], "engineering": [77], "that": [78, 96, 134], "expanded": [79], "simple": [81], "additional": [84], "contextual": [85], "information": [86], "static": [88], "program": [89], "analysis.": [90], "further": [94], "improvements": [95], "obtained": [98, 128], "fine-tuning": [100], "underlying": [102], "LLM.": [103], "The": [104], "fine": [105, 136], "tuning": [106, 170], "done": [108], "based": [109, 160], "on": [110, 161, 196], "custom": [112], "designed": [113], "template": [115], "has": [117], "pre-dependent": [118], "classes,": [119], "their": [120], "public": [121], "methods": [122, 163], "as": [123, 152], "well": [124], "two": [125], "exemplar": [126], "outputs": [127], "from": [129], "RAG.": [130], "establish": [133], "tuned": [137], "models": [138], "help": [139], "improve": [140], "correspondence": [142], "or": [143], "conformity": [144], "original": [147], "developer": [148], "written": [149], "measured": [153], "traditional": [156], "metrics": [157], "F1-score": [159], "used": [164], "generated": [167], "code.": [168], "Fine": [169], "8x7b": [173], "Mixture": [174], "Experts": [176], "(MoE)": [177], "model": [178, 189], "leads": [179], "to": [180, 193], "an": [181], "average": [182], "improvement": [183], "8\\%": [185], "over": [186], "base": [188], "comparable": [192], "scores": [195], "much": [198], "larger": [199], "8x22b": [200], "MoE": [201], "model.": [202]}, "primary_topic": {"id": "https://openalex.org/T10260", "display_name": "Software Engineering Research", "score": 0.8734999895095825, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10260", "display_name": "Software Engineering Research", "score": 0.8734999895095825, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10743", "display_name": "Software Testing and Debugging Techniques", "score": 0.121799997985363, "subfield": {"id": "https://openalex.org/subfields/1712", "display_name": "Software"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.002199999988079071, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.11006", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.11006", "pdf_url": "https://arxiv.org/pdf/2506.11006", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4407654582", "doi": "https://doi.org/10.1145/3718331", "title": "MixSong: Diverse and Strictly Formatted Chinese Poetry Generation", "publication_year": 2025, "publication_date": "2025-02-17", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Xinglong Song", "orcid": "https://orcid.org/0009-0005-1216-279X", "observed_orcids": ["https://orcid.org/0009-0005-1216-279X"]}, "institutions": [{"id": "https://openalex.org/I113940042", "display_name": "Shanghai University", "ror": "https://ror.org/006teas31", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I113940042"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Xinglong Song", "raw_affiliation_strings": ["Shanghai Film Academy, Shanghai University, Shanghai, China"], "raw_orcid": "https://orcid.org/0009-0005-1216-279X", "affiliations": [{"raw_affiliation_string": "Shanghai Film Academy, Shanghai University, Shanghai, China", "institution_ids": ["https://openalex.org/I113940042"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5116298293", "display_name": "Changlin Song", "orcid": "https://orcid.org/0009-0001-3501-7027", "observed_orcids": ["https://orcid.org/0009-0001-3501-7027"]}, "institutions": [{"id": "https://openalex.org/I113940042", "display_name": "Shanghai University", "ror": "https://ror.org/006teas31", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I113940042"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Changlin Song", "raw_affiliation_strings": ["Shanghai Film Academy, Shanghai University, Shanghai, China"], "raw_orcid": "https://orcid.org/0009-0001-3501-7027", "affiliations": [{"raw_affiliation_string": "Shanghai Film Academy, Shanghai University, Shanghai, China", "institution_ids": ["https://openalex.org/I113940042"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5116298294", "display_name": "Haolu Yu", "orcid": "https://orcid.org/0009-0004-8885-2407", "observed_orcids": ["https://orcid.org/0009-0004-8885-2407", "https://orcid.org/0009-0003-2434-4334"]}, "institutions": [{"id": "https://openalex.org/I113940042", "display_name": "Shanghai University", "ror": "https://ror.org/006teas31", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I113940042"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Haolu Yu", "raw_affiliation_strings": ["Shanghai Film Academy, Shanghai University, Shanghai, China"], "raw_orcid": "https://orcid.org/0009-0003-2434-4334", "affiliations": [{"raw_affiliation_string": "Shanghai Film Academy, Shanghai University, Shanghai, China", "institution_ids": ["https://openalex.org/I113940042"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5011407685", "display_name": "Yonghua Zhu", "orcid": "https://orcid.org/0000-0002-6419-9018", "observed_orcids": ["https://orcid.org/0000-0002-6419-9018"]}, "institutions": [{"id": "https://openalex.org/I113940042", "display_name": "Shanghai University", "ror": "https://ror.org/006teas31", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I113940042"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yonghua Zhu", "raw_affiliation_strings": ["Shanghai Film Academy, Shanghai University, Shanghai, China"], "raw_orcid": "https://orcid.org/0000-0002-6419-9018", "affiliations": [{"raw_affiliation_string": "Shanghai Film Academy, Shanghai University, Shanghai, China", "institution_ids": ["https://openalex.org/I113940042"]}]}, {"author_position": "last", "author": {"id": null, "display_name": "Hong Yao", "orcid": "https://orcid.org/0009-0000-7967-8962", "observed_orcids": ["https://orcid.org/0009-0000-7967-8962"]}, "institutions": [{"id": "https://openalex.org/I4210154361", "display_name": "Shanghai Jian Qiao University", "ror": "https://ror.org/04xdqtw10", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I4210154361"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Hong Yao", "raw_affiliation_strings": ["School of Information Technology, Shanghai Jianqiao University, Shanghai, China"], "raw_orcid": "https://orcid.org/0009-0000-7967-8962", "affiliations": [{"raw_affiliation_string": "School of Information Technology, Shanghai Jianqiao University, Shanghai, China", "institution_ids": ["https://openalex.org/I4210154361"]}]}], "abstract_inverted_index": {"Chinese": [0, 12], "poetry,": [1, 23], "renowned": [2], "for": [3, 144], "its": [4], "elegance": [5], "and": [6, 46, 62, 129, 139, 162, 171], "simplicity,": [7], "is": [8], "a": [9, 33, 52, 86], "hallmark": [10], "of": [11, 64, 89, 173], "culture.": [13], "While": [14], "neural": [15], "networks": [16], "have": [17], "made": [18], "significant": [19], "advancements": [20], "in": [21, 55, 108, 158, 168], "generating": [22, 145], "balancing": [24], "diversity": [25, 61, 170], "with": [26], "adherence": [27], "to": [28, 84, 96, 103], "rigid": [29], "structural": [30], "formats": [31], "remains": [32], "challenge.": [34], "Research": [35], "indicates": [36], "that": [37, 152], "factors": [38], "such": [39], "as": [40], "themes,": [41], "emotions": [42], "(e.g.,": [43, 48], "happiness,": [44], "sadness),": [45], "sentiments": [47], "positive,": [49], "negative)": [50], "play": [51], "crucial": [53], "role": [54], "poetic": [56], "creation,": [57], "influencing": [58], "both": [59, 159, 169], "the": [60, 65, 80, 101, 109, 141, 174], "quality": [63, 172], "generated": [66, 175], "content.": [67], "In": [68], "this": [69], "paper,": [70], "we": [71, 113], "propose": [72], "MixSong,": [73], "an": [74], "autoregressive": [75], "language": [76], "model": [77, 102], "based": [78], "on": [79], "Transformer": [81], "architecture,": [82], "designed": [83], "incorporate": [85], "wide": [87], "range": [88], "conditional": [90, 130], "factors.": [91], "MixSong": [92, 136, 153], "utilizes": [93], "adversarial": [94], "training": [95], "integrate": [97], "these": [98], "factors,": [99], "enabling": [100], "implicitly": [104], "learn": [105], "distributional": [106], "information": [107], "latent": [110], "space.": [111], "Additionally,": [112], "introduce": [114], "several": [115], "uniquely": [116], "customized": [117], "symbol": [118], "sets,": [119], "including": [120], "paragraph": [121], "identifiers,": [122, 124, 126, 128], "position": [123], "rhyme": [125], "tune": [127], "distinctive": [131], "identifiers.": [132], "These": [133], "symbols": [134], "help": [135], "effectively": [137], "capture": [138], "enforce": [140], "constraints": [142], "necessary": [143], "high-quality": [146], "poetry.": [147, 176], "Extensive": [148], "experimental": [149], "results": [150], "demonstrate": [151], "significantly": [154], "outperforms": [155], "existing": [156], "models": [157], "automatic": [160], "metrics": [161], "human": [163], "evaluations,": [164], "achieving": [165], "notable": [166], "improvements": [167]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9810000061988831, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9810000061988831, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10775", "display_name": "Generative Adversarial Networks and Image Synthesis", "score": 0.007899999618530273, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.003700000001117587, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1145/3718331", "is_oa": false, "landing_page_url": "https://doi.org/10.1145/3718331", "pdf_url": null, "source": {"id": "https://openalex.org/S4306421405", "display_name": "ACM Transactions on Asian and Low-Resource Language Information Processing", "issn_l": "2375-4699", "issn": ["2375-4699", "2375-4702"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-1", "norway-1"], "host_organization": "https://openalex.org/P4310319798", "host_organization_name": "Association for Computing Machinery", "host_organization_lineage": ["https://openalex.org/P4310319798"], "host_organization_lineage_names": ["Association for Computing Machinery"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "ACM Transactions on Asian and Low-Resource Language Information Processing", "raw_type": "journal-article"}, "cited_by_count": 2, "is_retracted": false}, {"id": "https://openalex.org/W4413378215", "doi": "https://doi.org/10.2196/preprints.82400", "title": "Evaluating Online Medical Information Quality in Colon Surgery Using Pre-Tuned Large Language Models. (Preprint)", "publication_year": 2025, "publication_date": "2025-08-14", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5088411115", "display_name": "Vincent Ochs", "orcid": "https://orcid.org/0000-0003-0237-7496", "observed_orcids": ["https://orcid.org/0000-0003-0237-7496"]}, "institutions": [{"id": "https://openalex.org/I1850255", "display_name": "University of Basel", "ror": "https://ror.org/02s6k3f65", "country_code": "CH", "type": "education", "lineage": ["https://openalex.org/I1850255"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Vincent Ochs", "raw_affiliation_strings": [". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "Department of Biomedical Engineering University of Basel Allschwil CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "institution_ids": ["https://openalex.org/I1850255"]}, {"raw_affiliation_string": "Department of Biomedical Engineering University of Basel Allschwil CH", "institution_ids": ["https://openalex.org/I1850255"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119370361", "display_name": "S\u00e9bastien Muheim", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I1850255", "display_name": "University of Basel", "ror": "https://ror.org/02s6k3f65", "country_code": "CH", "type": "education", "lineage": ["https://openalex.org/I1850255"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "S\u00e9bastien Muheim", "raw_affiliation_strings": [". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "Department of Biomedical Engineering University of Basel Allschwil CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "institution_ids": ["https://openalex.org/I1850255"]}, {"raw_affiliation_string": "Department of Biomedical Engineering University of Basel Allschwil CH", "institution_ids": ["https://openalex.org/I1850255"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5037902862", "display_name": "Florian Ponholzer", "orcid": "https://orcid.org/0000-0003-3486-7166", "observed_orcids": ["https://orcid.org/0000-0003-3486-7166"]}, "institutions": [{"id": "https://openalex.org/I143397708", "display_name": "Innsbruck Medical University", "ror": "https://ror.org/03pt86f80", "country_code": "AT", "type": "education", "lineage": ["https://openalex.org/I143397708"]}, {"id": "https://openalex.org/I190249584", "display_name": "Universit\u00e4t Innsbruck", "ror": "https://ror.org/054pv6659", "country_code": "AT", "type": "education", "lineage": ["https://openalex.org/I190249584"]}], "countries": ["AT"], "is_corresponding": false, "raw_author_name": "Florian Ponholzer", "raw_affiliation_strings": [". Department of Visceral, Transplant and Thoracic Surgery , Center of Operative Medicine , Medical University of Innsbruck , 6020 Innsbruck , Austria", "Department of Visceral , Transplant and Thoracic Surgery Center of Operative Medicine Medical University of Innsbruck Innsbruck AT"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Visceral, Transplant and Thoracic Surgery , Center of Operative Medicine , Medical University of Innsbruck , 6020 Innsbruck , Austria", "institution_ids": ["https://openalex.org/I143397708", "https://openalex.org/I190249584"]}, {"raw_affiliation_string": "Department of Visceral , Transplant and Thoracic Surgery Center of Operative Medicine Medical University of Innsbruck Innsbruck AT", "institution_ids": ["https://openalex.org/I143397708", "https://openalex.org/I190249584"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5018522046", "display_name": "Dietmar \u00d6fner", "orcid": "https://orcid.org/0000-0001-8909-8566", "observed_orcids": ["https://orcid.org/0000-0001-8909-8566"]}, "institutions": [{"id": "https://openalex.org/I143397708", "display_name": "Innsbruck Medical University", "ror": "https://ror.org/03pt86f80", "country_code": "AT", "type": "education", "lineage": ["https://openalex.org/I143397708"]}, {"id": "https://openalex.org/I190249584", "display_name": "Universit\u00e4t Innsbruck", "ror": "https://ror.org/054pv6659", "country_code": "AT", "type": "education", "lineage": ["https://openalex.org/I190249584"]}], "countries": ["AT"], "is_corresponding": false, "raw_author_name": "Dietmar \u00d6fner", "raw_affiliation_strings": [". Department of Visceral, Transplant and Thoracic Surgery , Center of Operative Medicine , Medical University of Innsbruck , 6020 Innsbruck , Austria", "Department of Visceral , Transplant and Thoracic Surgery Center of Operative Medicine Medical University of Innsbruck Innsbruck AT"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Visceral, Transplant and Thoracic Surgery , Center of Operative Medicine , Medical University of Innsbruck , 6020 Innsbruck , Austria", "institution_ids": ["https://openalex.org/I143397708", "https://openalex.org/I190249584"]}, {"raw_affiliation_string": "Department of Visceral , Transplant and Thoracic Surgery Center of Operative Medicine Medical University of Innsbruck Innsbruck AT", "institution_ids": ["https://openalex.org/I143397708", "https://openalex.org/I190249584"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5059272660", "display_name": "Antonio J. Rodr\u00edguez-S\u00e1nchez", "orcid": "https://orcid.org/0000-0002-3264-5060", "observed_orcids": ["https://orcid.org/0000-0002-3264-5060"]}, "institutions": [{"id": "https://openalex.org/I190249584", "display_name": "Universit\u00e4t Innsbruck", "ror": "https://ror.org/054pv6659", "country_code": "AT", "type": "education", "lineage": ["https://openalex.org/I190249584"]}], "countries": ["AT"], "is_corresponding": false, "raw_author_name": "Antonio J. Rodr\u00edguez-S\u00e1nchez", "raw_affiliation_strings": [". Intelligent and Interactive Systems , Department of Computer Science , University of Innsbruck , Innsbruck , Austria", "Department of Computer Science, Intelligent and Interactive Systems University of Innsbruck Innsbruck AT"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Intelligent and Interactive Systems , Department of Computer Science , University of Innsbruck , Innsbruck , Austria", "institution_ids": ["https://openalex.org/I190249584"]}, {"raw_affiliation_string": "Department of Computer Science, Intelligent and Interactive Systems University of Innsbruck Innsbruck AT", "institution_ids": ["https://openalex.org/I190249584"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5071493577", "display_name": "Mariana Flaifel", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I165862685", "display_name": "St George's, University of London", "ror": "https://ror.org/040f08y74", "country_code": "GB", "type": "education", "lineage": ["https://openalex.org/I124357947", "https://openalex.org/I165862685", "https://openalex.org/I4401726869"]}], "countries": ["GB"], "is_corresponding": false, "raw_author_name": "Mariana Flaifel", "raw_affiliation_strings": [". Department of Surgery , St. George's University of London , London , UK", "Department of Surgery , St. George's University of London London GB"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Surgery , St. George's University of London , London , UK", "institution_ids": ["https://openalex.org/I165862685"]}, {"raw_affiliation_string": "Department of Surgery , St. George's University of London London GB", "institution_ids": ["https://openalex.org/I165862685"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100775611", "display_name": "Amit Kumar", "orcid": "https://orcid.org/0000-0001-5920-0048", "observed_orcids": ["https://orcid.org/0000-0001-5920-0048"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Amit Kumar", "raw_affiliation_strings": [". Medoc , Data Since and Statistic , Basel , Switzerland", "Data Since and Statistic Medoc Basel CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Medoc , Data Since and Statistic , Basel , Switzerland", "institution_ids": []}, {"raw_affiliation_string": "Data Since and Statistic Medoc Basel CH", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5033937723", "display_name": "Julia Wolleb", "orcid": "https://orcid.org/0000-0003-4087-5920", "observed_orcids": ["https://orcid.org/0000-0003-4087-5920"]}, "institutions": [{"id": "https://openalex.org/I1850255", "display_name": "University of Basel", "ror": "https://ror.org/02s6k3f65", "country_code": "CH", "type": "education", "lineage": ["https://openalex.org/I1850255"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Julia Wolleb", "raw_affiliation_strings": [". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "Department of Biomedical Engineering University of Basel Allschwil CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "institution_ids": ["https://openalex.org/I1850255"]}, {"raw_affiliation_string": "Department of Biomedical Engineering University of Basel Allschwil CH", "institution_ids": ["https://openalex.org/I1850255"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5031248007", "display_name": "Florentin Bieder", "orcid": "https://orcid.org/0000-0001-9558-0623", "observed_orcids": ["https://orcid.org/0000-0001-9558-0623"]}, "institutions": [{"id": "https://openalex.org/I1850255", "display_name": "University of Basel", "ror": "https://ror.org/02s6k3f65", "country_code": "CH", "type": "education", "lineage": ["https://openalex.org/I1850255"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Florentin Bieder", "raw_affiliation_strings": [". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "Department of Biomedical Engineering University of Basel Allschwil CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "institution_ids": ["https://openalex.org/I1850255"]}, {"raw_affiliation_string": "Department of Biomedical Engineering University of Basel Allschwil CH", "institution_ids": ["https://openalex.org/I1850255"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5029548133", "display_name": "Stephanie Taha\u2010Mehlitz", "orcid": "https://orcid.org/0000-0002-0962-6275", "observed_orcids": ["https://orcid.org/0000-0002-0962-6275"]}, "institutions": [{"id": "https://openalex.org/I2802542264", "display_name": "University Hospital of Basel", "ror": "https://ror.org/04k51q396", "country_code": "CH", "type": "healthcare", "lineage": ["https://openalex.org/I2802542264"]}, {"id": "https://openalex.org/I4210135747", "display_name": "St. Claraspital", "ror": "https://ror.org/04ahnxd67", "country_code": "CH", "type": "healthcare", "lineage": ["https://openalex.org/I4210135747"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Stephanie Taha-Mehlitz", "raw_affiliation_strings": [". Clarunis , Department of Visceral Surgery , University Center for Gastrointestinal and Liver Diseases , St. Clara Hospital and University Hospital , Basel , Switzerland", "Department of Visceral Surgery , University Center for Gastrointestinal and Liver Diseases St. Clara Hospital and University Hospital Basel CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Clarunis , Department of Visceral Surgery , University Center for Gastrointestinal and Liver Diseases , St. Clara Hospital and University Hospital , Basel , Switzerland", "institution_ids": ["https://openalex.org/I2802542264", "https://openalex.org/I4210135747"]}, {"raw_affiliation_string": "Department of Visceral Surgery , University Center for Gastrointestinal and Liver Diseases St. Clara Hospital and University Hospital Basel CH", "institution_ids": ["https://openalex.org/I2802542264", "https://openalex.org/I4210135747"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5027401046", "display_name": "Michael D. Honaker", "orcid": "https://orcid.org/0000-0003-4574-2094", "observed_orcids": ["https://orcid.org/0000-0003-4574-2094"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Michael Drew Honaker", "raw_affiliation_strings": ["Department of Biomedical Engineering ,"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Biomedical Engineering ,", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5079233267", "display_name": "ROBERT D. ROSENBERG", "orcid": "https://orcid.org/0000-0003-1849-0709", "observed_orcids": ["https://orcid.org/0000-0003-1849-0709"]}, "institutions": [{"id": "https://openalex.org/I4210091012", "display_name": "Kantonsspital Baselland", "ror": "https://ror.org/00b747122", "country_code": "CH", "type": "healthcare", "lineage": ["https://openalex.org/I4210091012"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Robert Rosenberg", "raw_affiliation_strings": [". Department of Visceral Surgery , Cantonal Hospital Basel-Land , Liestal , Switzerland", "Department of Visceral Surgery Cantonal Hospital Basel-Land Liestal CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Visceral Surgery , Cantonal Hospital Basel-Land , Liestal , Switzerland", "institution_ids": ["https://openalex.org/I4210091012"]}, {"raw_affiliation_string": "Department of Visceral Surgery Cantonal Hospital Basel-Land Liestal CH", "institution_ids": ["https://openalex.org/I4210091012"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5070616013", "display_name": "Philippe C. Cattin", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I1850255", "display_name": "University of Basel", "ror": "https://ror.org/02s6k3f65", "country_code": "CH", "type": "education", "lineage": ["https://openalex.org/I1850255"]}], "countries": ["CH"], "is_corresponding": false, "raw_author_name": "Philippe C. Cattin", "raw_affiliation_strings": [". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "Department of Biomedical Engineering University of Basel Allschwil CH"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": ". Department of Biomedical Engineering , Faculty of Medicine , University of Basel , 4123 Allschwil , Switzerland", "institution_ids": ["https://openalex.org/I1850255"]}, {"raw_affiliation_string": "Department of Biomedical Engineering University of Basel Allschwil CH", "institution_ids": ["https://openalex.org/I1850255"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5033033571", "display_name": "Anas Taha", "orcid": "https://orcid.org/0000-0001-8100-1598", "observed_orcids": ["https://orcid.org/0000-0001-8100-1598"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Anas Taha", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"BACKGROUND": [0], "Online": [1], "medical": [2, 44, 348, 378], "information": [3, 92], "in": [4, 8, 32, 127, 154, 341, 407], "colon": [5, 95, 123], "surgery,": [6], "particularly": [7, 332], "robotic-assisted": [9], "procedures,": [10], "is": [11, 357], "frequently": [12], "inconsistent,": [13], "outdated,": [14], "or": [15, 108], "misleading,": [16], "which": [17], "can": [18], "hinder": [19], "patient": [20, 389], "education": [21], "and": [22, 74, 103, 136, 148, 194, 197, 206, 241, 266, 385, 399, 402], "informed": [23], "decision-making.": [24], "Large": [25], "Language": [26], "Models": [27, 180], "(LLMs)": [28], "have": [29], "shown": [30], "promise": [31], "complex": [33], "text": [34], "analysis,": [35], "but": [36, 69, 296, 354], "their": [37], "potential": [38], "for": [39, 56, 65, 114, 316, 329, 377, 405], "assessing": [40], "the": [41, 80, 88, 99, 248, 289, 306, 368, 383], "quality": [42, 89, 344, 356], "of": [43, 82, 90, 346, 370, 387], "web": [45, 119], "content": [46, 379], "has": [47], "not": [48], "been": [49], "systematically": [50, 78], "explored.": [51], "The": [52], "Ensuring": [53], "Quality": [54], "Information": [55], "Patients": [57], "(EQIP)": [58], "36": [59, 101], "criteria": [60], "offer": [61], "a": [62, 209, 358], "standardized": [63], "framework": [64], "evaluating": [66], "patient-facing": [67], "materials,": [68], "manual": [70], "assessments": [71], "are": [72, 334], "time-consuming": [73], "resource-intensive.": [75], "OBJECTIVE": [76], "To": [77], "evaluate": [79], "ability": [81], "multiple": [83], "transformer-based": [84], "LLMs": [85, 372], "to": [86, 98, 104, 164], "classify": [87], "online": [91, 347, 388], "on": [93, 121, 133], "robotic": [94, 122], "surgery": [96, 124], "according": [97], "EQIP": [100, 161, 216], "criteria,": [102], "determine": [105], "whether": [106], "embedding-oriented": [107], "generative": [109, 323, 339], "models": [110, 327, 340], "provide": [111], "better": [112], "performance": [113, 250, 308], "this": [115], "task.": [116], "METHODS": [117], "English-language": [118], "pages": [120], "were": [125, 151, 169, 201, 224], "collected": [126], "December": [128], "2023": [129], "via": [130, 208, 226], "automated": [131, 342], "searches": [132], "Google,": [134], "Bing,": [135], "Yahoo": [137], "using": [138, 232], "predefined": [139], "keywords.": [140], "From": [141], "450": [142], "retrieved": [143], "URLs,": [144], "duplicates,": [145], "irrelevant": [146], "content,": [147], "professional/academic": [149], "sites": [150], "removed,": [152], "resulting": [153], "47": [155], "unique": [156], "patient-targeted": [157], "websites": [158], "with": [159, 236, 282], "expert": [160], "ratings.": [162], "Due": [163], "input": [165], "length": [166], "limitations,": [167], "texts": [168], "split": [170], "into": [171, 203, 373], "overlapping": [172], "512-token": [173], "chunks": [174], "(50-token": [175], "overlap)": [176], "before": [177], "embedding": [178, 291, 330, 355], "extraction.": [179], "evaluated": [181], "included": [182], "base": [183], "BERT,": [184, 187], "custom": [185, 267], "pre-trained": [186, 189, 268], "BiomedBERT,": [188, 190], "SBERT,": [191, 333], "LLaMA-2": [192, 281], "(concatenated": [193], "averaged": [195, 290], "embeddings),": [196], "Longformer.": [198], "Chunk": [199], "embeddings": [200, 205, 284, 320], "concatenated": [202, 283], "document-level": [204], "classified": [207], "feedforward": [210], "multilabel": [211], "neural": [212], "network": [213], "predicting": [214], "19": [215], "items.": [217], "Hyperparameters": [218], "(batch": [219], "size,": [220], "learning": [221], "rate,": [222], "dropout)": [223], "optimized": [225, 315, 328], "grid": [227], "search.": [228], "Performance": [229], "was": [230], "assessed": [231], "stratified": [233], "5-fold": [234], "cross-validation,": [235], "accuracy,": [237], "precision,": [238], "recall,": [239], "F1-score,": [240], "ROC-AUC": [242], "as": [243], "metrics.": [244], "RESULTS": [245], "SBERT": [246], "achieved": [247], "highest": [249], "(F1-score": [251], "=": [252, 255, 258, 286, 294, 310], "80.16,": [253], "accuracy": [254], "79.05%,": [256], "AUC": [257], "0.79),": [259, 275], "outperforming": [260], "all": [261], "other": [262], "models.": [263], "Pre-trained": [264], "BiomedBERT": [265], "BERT": [269], "performed": [270], "comparably": [271], "well": [272], "(F1": [273, 285, 293, 309], "\u2248": [274], "indicating": [276], "benefits": [277], "from": [278], "domain-specific": [279], "adaptation.": [280], "0.784)": [287], "outperformed": [288], "version": [292], "0.765)": [295], "remained": [297], "below": [298], "SBERT.": [299], "Longformer,": [300], "despite": [301], "handling": [302], "longer": [303], "contexts,": [304], "had": [305], "lowest": [307], "70.07).": [311], "Across": [312], "models,": [313], "those": [314], "generating": [317], "semantically": [318], "meaningful": [319], "consistently": [321], "surpassed": [322], "architectures.": [324], "CONCLUSIONS": [325], "Transformer-based": [326], "generation,": [331], "more": [335, 359], "effective": [336], "than": [337, 362], "large": [338], "EQIP-based": [343], "assessment": [345], "content.": [349], "Domain-specific": [350], "adaptation": [351], "improves": [352], "performance,": [353], "critical": [360], "determinant": [361], "model": [363], "size.": [364], "These": [365], "findings": [366], "support": [367], "integration": [369], "embedding-focused": [371], "scalable,": [374], "AI-driven": [375], "tools": [376], "evaluation,": [380], "potentially": [381], "enhancing": [382], "accessibility": [384], "reliability": [386], "education.": [390], "Future": [391], "research": [392], "should": [393], "expand": [394], "datasets,": [395], "explore": [396], "ensemble": [397], "approaches,": [398], "address": [400], "regulatory": [401], "ethical": [403], "considerations": [404], "deployment": [406], "healthcare": [408], "contexts.": [409]}, "primary_topic": {"id": "https://openalex.org/T10737", "display_name": "Health Literacy and Information Accessibility", "score": 0.9491999745368958, "subfield": {"id": "https://openalex.org/subfields/3600", "display_name": "General Health Professions"}, "field": {"id": "https://openalex.org/fields/36", "display_name": "Health Professions"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}, "topics": [{"id": "https://openalex.org/T10737", "display_name": "Health Literacy and Information Accessibility", "score": 0.9491999745368958, "subfield": {"id": "https://openalex.org/subfields/3600", "display_name": "General Health Professions"}, "field": {"id": "https://openalex.org/fields/36", "display_name": "Health Professions"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.019200000911951065, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11636", "display_name": "Artificial Intelligence in Healthcare and Education", "score": 0.007699999958276749, "subfield": {"id": "https://openalex.org/subfields/2718", "display_name": "Health Informatics"}, "field": {"id": "https://openalex.org/fields/27", "display_name": "Medicine"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}], "primary_location": {"id": "doi:10.2196/preprints.82400", "is_oa": true, "landing_page_url": "https://doi.org/10.2196/preprints.82400", "pdf_url": "https://s3.ca-central-1.amazonaws.com/assets.jmir.org/assets/preprints/preprint-82400-submitted.pdf", "source": null, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414625741", "doi": "https://doi.org/10.48550/arxiv.2506.17671", "title": "TPTT: Transforming Pretrained Transformers into Titans", "publication_year": 2025, "publication_date": "2025-06-21", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5045207062", "display_name": "Fabien Furfaro", "orcid": "https://orcid.org/0000-0003-1997-1141", "observed_orcids": ["https://orcid.org/0000-0003-1997-1141"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Furfaro, Fabien", "raw_affiliation_strings": ["Chercheur ind\u00e9pendant (France)"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Chercheur ind\u00e9pendant (France)", "institution_ids": []}]}], "abstract_inverted_index": {"Transformer-based": [0], "large": [1], "language": [2, 12], "models": [3, 114, 208], "(LLMs)": [4], "have": [5], "achieved": [6], "strong": [7], "performance": [8], "across": [9, 107], "many": [10], "natural": [11], "processing": [13], "tasks.": [14], "Nonetheless,": [15], "their": [16], "quadratic": [17], "computational": [18, 185], "and": [19, 34, 59, 77, 100, 132, 209, 222], "memory": [20, 61], "requirements,": [21], "particularly": [22], "in": [23, 37, 102, 129, 148, 152], "self-attention": [24], "layers,": [25], "pose": [26], "challenges": [27], "for": [28, 35, 198], "efficient": [29], "inference": [30], "on": [31, 90, 113, 122, 206], "long": [32], "contexts": [33], "deployment": [36], "resource-limited": [38], "environments.": [39], "We": [40, 87], "present": [41], "TPTT": [42, 72, 89, 192], "(Transforming": [43], "Pretrained": [44], "Transformers": [45, 54], "into": [46, 168], "Titans),": [47], "a": [48, 144, 165, 169, 210], "framework": [49], "designed": [50], "to": [51, 104, 135, 143, 163, 218], "augment": [52], "pretrained": [53, 92, 196], "with": [55, 79, 115, 183, 201], "linearized": [56], "attention": [57], "(LiZA)": [58], "internal": [60], "gating": [62], "via": [63], "Memory": [64], "as": [65, 83], "Gate": [66], "(MaG),": [67], "applied": [68], "without": [69], "full": [70], "retraining.": [71], "supports": [73], "parameter-efficient": [74], "fine-tuning": [75], "(LoRA)": [76], "integrates": [78], "standard": [80], "toolkits": [81], "such": [82], "Hugging": [84], "Face": [85], "Transformers.": [86], "evaluated": [88, 120], "several": [91], "models,": [93], "including": [94], "Llama-1B,": [95], "OlMoE-1B-7B,": [96], "Qwen2.5-1.5B,": [97], "Gemma3-270m,": [98], "OpenELM-1.3B,": [99], "Mistral-7B,": [101], "order": [103], "assess": [105], "applicability": [106], "architectures": [108], "of": [109, 213, 224], "different": [110], "scales.": [111], "Experiments": [112], "approximately": [116], "1": [117], "billion": [118], "parameters,": [119], "primarily": [121], "the": [123, 174, 220, 225], "MMLU": [124], "benchmark,": [125], "suggest": [126], "potential": [127], "improvements": [128], "both": [130], "efficiency": [131], "accuracy": [133], "compared": [134], "baseline": [136], "models.": [137], "For": [138], "example,": [139], "Titans-Llama-1B": [140], "exhibited": [141], "up": [142], "20\\%": [145], "relative": [146], "increase": [147], "Exact": [149], "Match": [150], "scores": [151], "one-shot": [153], "evaluation.": [154], "An": [155], "additional": [156], "finding": [157], "is": [158, 161, 228], "that": [159, 191], "it": [160], "possible": [162], "convert": [164], "quadratic-attention": [166], "model": [167, 172], "purely": [170], "linear-attention": [171], "using": [173], "DeltaProduct": [175], "mechanism.": [176], "All": [177], "training": [178], "runs": [179], "were": [180], "carried": [181], "out": [182], "modest": [184], "resources.": [186], "These": [187], "preliminary": [188], "findings": [189], "indicate": [190], "may": [193], "help": [194], "adapt": [195], "LLMs": [197], "long-context": [199], "tasks": [200], "limited": [202], "overhead.": [203], "Further": [204], "studies": [205], "larger": [207], "broader": [211], "set": [212], "benchmarks": [214], "will": [215], "be": [216], "necessary": [217], "evaluate": [219], "generality": [221], "robustness": [223], "framework.": [226], "Code": [227], "available": [229], "at": [230, 235], "https://github.com/fabienfrfr/tptt": [231], ".": [232, 237], "Python": [233], "package": [234], "https://pypi.org/project/tptt/": [236]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9096999764442444, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9096999764442444, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.0575999990105629, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.014800000004470348, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2506.17671", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2506.17671", "pdf_url": "https://arxiv.org/pdf/2506.17671", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414946794", "doi": "https://doi.org/10.48550/arxiv.2510.04568", "title": "COSMIR: Chain Orchestrated Structured Memory for Iterative Reasoning over Long Context", "publication_year": 2025, "publication_date": "2025-10-06", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5107933712", "display_name": "Naman Gupta", "orcid": "https://orcid.org/0009-0000-4198-4892", "observed_orcids": ["https://orcid.org/0009-0000-4198-4892"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gupta, Naman", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5119897630", "display_name": "Shreeyash Gowaikar", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gowaikar, Shreeyash", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Iyer, Arun", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Iyer, Arun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5051051894", "display_name": "Kirankumar Shiragur", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Shiragur, Kirankumar", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5063416855", "display_name": "Ramakrishna Bairi", "orcid": "https://orcid.org/0009-0009-9191-3748", "observed_orcids": ["https://orcid.org/0009-0009-9191-3748"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bairi, Ramakrishna B", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Maurya, Rishikesh", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Maurya, Rishikesh", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Maiti, Ritabrata", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Maiti, Ritabrata", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5026321774", "display_name": "Sankarshan Damle", "orcid": "https://orcid.org/0000-0003-1460-6102", "observed_orcids": ["https://orcid.org/0000-0003-1460-6102"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Damle, Sankarshan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5032355009", "display_name": "Shachee Mishra Gupta", "orcid": "https://orcid.org/0000-0003-1332-690X", "observed_orcids": ["https://orcid.org/0000-0003-1332-690X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gupta, Shachee Mishra", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Reasoning": [0], "over": [1, 164], "very": [2], "long": [3], "inputs": [4], "remains": [5], "difficult": [6], "for": [7, 65], "large": [8], "language": [9], "models": [10], "(LLMs).": [11], "Common": [12], "workarounds": [13], "either": [14], "shrink": [15], "the": [16, 24, 107, 115, 120, 130, 136, 153], "input": [17], "via": [18, 96], "retrieval": [19], "(risking": [20], "missed": [21], "evidence),": [22], "enlarge": [23], "context": [25], "window": [26], "(straining": [27], "selectivity),": [28], "or": [29], "stage": [30], "multiple": [31], "agents": [32, 49, 93], "to": [33, 106], "read": [34], "in": [35], "pieces.": [36], "In": [37], "staged": [38], "pipelines": [39], "(e.g.,": [40], "Chain": [41], "of": [42], "Agents,": [43], "CoA),": [44], "free-form": [45], "summaries": [46], "passed": [47], "between": [48], "can": [50], "discard": [51], "crucial": [52], "details": [53], "and": [54, 135, 147, 161], "amplify": [55], "early": [56], "mistakes.": [57], "We": [58], "introduce": [59], "COSMIR": [60, 156], "(Chain": [61], "Orchestrated": [62], "Structured": [63], "Memory": [64], "Iterative": [66], "Reasoning),": [67], "a": [68, 77, 85, 97, 165], "chain-style": [69], "framework": [70], "that": [71], "replaces": [72], "ad": [73], "hoc": [74], "messages": [75], "with": [76], "structured": [78], "memory.": [79, 109, 121], "A": [80, 110], "Planner": [81], "agent": [82, 112], "first": [83], "turns": [84], "user": [86], "query": [87], "into": [88], "concrete,": [89], "checkable": [90], "sub-questions.": [91], "worker": [92, 137], "process": [94], "chunks": [95], "fixed": [98], "micro-cycle:": [99], "Extract,": [100], "Infer,": [101], "Refine,": [102], "writing": [103], "all": [104], "updates": [105], "shared": [108], "Manager": [111], "then": [113], "Synthesizes": [114], "final": [116], "answer": [117], "directly": [118], "from": [119, 152], "This": [122], "preserves": [123], "step-wise": [124], "read-then-reason": [125], "benefits": [126], "while": [127], "changing": [128], "both": [129], "communication": [131], "medium": [132], "(structured": [133], "memory)": [134], "procedure": [138], "(fixed": [139], "micro-cycle),": [140], "yielding": [141], "higher": [142], "faithfulness,": [143], "better": [144], "long-range": [145], "aggregation,": [146], "auditability.": [148], "On": [149], "long-context": [150], "QA": [151], "HELMET": [154], "suite,": [155], "reduces": [157], "propagation-stage": [158], "information": [159], "loss": [160], "improves": [162], "accuracy": [163], "CoA": [166], "baseline.": [167]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8070999979972839, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.8070999979972839, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.06019999831914902, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.046300001442432404, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2510.04568", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2510.04568", "pdf_url": "https://arxiv.org/pdf/2510.04568", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417467409", "doi": "https://doi.org/10.48550/arxiv.2512.14244", "title": "From Context to EDUs: Faithful and Structured Context Compression via Elementary Discourse Unit Decomposition", "publication_year": 2025, "publication_date": "2025-12-16", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Zhou, Yiqing", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I139759216", "display_name": "Beijing University of Posts and Telecommunications", "ror": "https://ror.org/04w9fbh59", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I139759216"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zhou, Yiqing", "raw_affiliation_strings": ["Beijing University of Posts and Telecommunications \u22c6"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing University of Posts and Telecommunications \u22c6", "institution_ids": ["https://openalex.org/I139759216"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100736935", "display_name": "Lei Yu", "orcid": "https://orcid.org/0000-0002-7329-4631", "observed_orcids": ["https://orcid.org/0000-0002-7329-4631"]}, "institutions": [{"id": "https://openalex.org/I139759216", "display_name": "Beijing University of Posts and Telecommunications", "ror": "https://ror.org/04w9fbh59", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I139759216"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Lei, Yu", "raw_affiliation_strings": ["Beijing University of Posts and Telecommunications \u22c6"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Beijing University of Posts and Telecommunications \u22c6", "institution_ids": ["https://openalex.org/I139759216"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5022397026", "display_name": "Shuzheng Si", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Si, Shuzheng", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5031777966", "display_name": "Qingyan Sun", "orcid": "https://orcid.org/0000-0003-3219-4332", "observed_orcids": ["https://orcid.org/0000-0003-3219-4332"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Sun, Qingyan", "raw_affiliation_strings": ["Department of Computer Science and Technology , Tsinghua University \u2662"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science and Technology , Tsinghua University \u2662", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Wang, Wei", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Wang, Wei", "raw_affiliation_strings": ["Department of Computer Science and Technology , Tsinghua University \u2662"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science and Technology , Tsinghua University \u2662", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Wu, Yifei", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Yifei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5045918129", "display_name": "Hao Wen", "orcid": "https://orcid.org/0000-0002-0060-2840", "observed_orcids": ["https://orcid.org/0000-0002-0060-2840", "https://orcid.org/0009-0008-6279-9440", "https://orcid.org/0009-0008-8450-7795"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wen, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5100389273", "display_name": "Gang Chen", "orcid": "https://orcid.org/0000-0002-5194-8956", "observed_orcids": ["https://orcid.org/0000-0002-5194-8956"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chen, Gang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5030183493", "display_name": "Fanchao Qi", "orcid": "https://orcid.org/0000-0002-4400-4033", "observed_orcids": ["https://orcid.org/0000-0002-4400-4033"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Qi, Fanchao", "raw_affiliation_strings": ["Department of Computer Science and Technology , Tsinghua University \u2662"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science and Technology , Tsinghua University \u2662", "institution_ids": ["https://openalex.org/I99065089"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5046448314", "display_name": "Maosong Sun", "orcid": "https://orcid.org/0000-0002-6011-6115", "observed_orcids": ["https://orcid.org/0000-0002-6011-6115"]}, "institutions": [{"id": "https://openalex.org/I99065089", "display_name": "Tsinghua University", "ror": "https://ror.org/03cve4549", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I99065089"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Sun, Maosong", "raw_affiliation_strings": ["Department of Computer Science and Technology , Tsinghua University \u2662"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Computer Science and Technology , Tsinghua University \u2662", "institution_ids": ["https://openalex.org/I99065089"]}]}], "abstract_inverted_index": {"Managing": [0], "extensive": [1], "context": [2, 86], "remains": [3], "a": [4, 69, 89, 99, 119, 136], "critical": [5], "bottleneck": [6], "for": [7, 126], "Large": [8], "Language": [9], "Models": [10], "(LLMs),": [11], "particularly": [12], "in": [13], "applications": [14], "like": [15], "long-document": [16], "question": [17], "answering": [18], "and": [19, 29, 54, 80, 155], "autonomous": [20], "agents": [21], "where": [22], "lengthy": [23], "inputs": [24], "incur": [25], "high": [26], "computational": [27], "costs": [28], "introduce": [30, 64], "noise.": [31], "Existing": [32], "compression": [33, 72, 87, 166], "techniques": [34], "often": [35], "disrupt": [36], "local": [37], "coherence": [38], "through": [39], "discrete": [40], "token": [41], "removal": [42], "or": [43], "rely": [44], "on": [45], "implicit": [46], "latent": [47], "encoding": [48], "that": [49, 147], "suffers": [50], "from": [51, 174], "positional": [52], "bias": [53], "incompatibility": [55], "with": [56], "closed-source": [57], "APIs.": [58], "To": [59, 128], "address": [60], "these": [61], "limitations,": [62], "we": [63, 133], "the": [65], "EDU-based": [66], "Context": [67], "Compressor,": [68], "novel": [70], "explicit": [71], "framework": [73], "designed": [74], "to": [75, 112, 115, 177], "preserve": [76], "both": [77], "global": [78], "structure": [79], "fine-grained": [81], "details.": [82], "Our": [83], "approach": [84], "reformulates": [85], "as": [88], "structure-then-select": [90], "process.": [91], "First,": [92], "our": [93, 148, 164], "LingoEDU": [94], "transforms": [95], "linear": [96], "text": [97], "into": [98], "structural": [100, 131, 152], "relation": [101], "tree": [102], "of": [103, 140], "Elementary": [104], "Discourse": [105], "Units": [106], "(EDUs)": [107], "which": [108], "are": [109], "anchored": [110], "strictly": [111], "source": [113], "indices": [114], "eliminate": [116], "hallucination.": [117], "Second,": [118], "lightweight": [120], "ranking": [121], "module": [122], "selects": [123], "query-relevant": [124], "sub-trees": [125], "linearization.": [127], "rigorously": [129], "evaluate": [130], "understanding,": [132], "release": [134], "StructBench,": [135], "manually": [137], "annotated": [138], "dataset": [139], "248": [141], "diverse": [142], "documents.": [143], "Empirical": [144], "results": [145], "demonstrate": [146], "method": [149], "achieves": [150], "state-of-the-art": [151], "prediction": [153], "accuracy": [154], "significantly": [156], "outperforms": [157], "frontier": [158], "LLMs": [159], "while": [160], "reducing": [161], "costs.": [162], "Furthermore,": [163], "structure-aware": [165], "substantially": [167], "enhances": [168], "performance": [169], "across": [170], "downstream": [171], "tasks": [172, 176], "ranging": [173], "long-context": [175], "complex": [178], "Deep": [179], "Search": [180], "scenarios.": [181]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9541000127792358, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9541000127792358, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.02969999983906746, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.004600000102072954, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2512.14244", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2512.14244", "pdf_url": "https://arxiv.org/pdf/2512.14244", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4408561719", "doi": "https://doi.org/10.31219/osf.io/b94ng_v1", "title": "Long Text Language Ensemble Models: Extension of Retrieval-Augmented Generation (RAG)", "publication_year": 2025, "publication_date": "2025-03-18", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5014256995", "display_name": "Abiodun Christian Ibiloye", "orcid": "https://orcid.org/0000-0001-9404-9830", "observed_orcids": ["https://orcid.org/0000-0001-9404-9830"]}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Abiodun Christian Ibiloye", "raw_affiliation_strings": ["Department of Higher Education State Ministry of Education , Kaduna Kaduna State , Nigeria"], "raw_orcid": "https://orcid.org/0000-0001-9404-9830", "affiliations": [{"raw_affiliation_string": "Department of Higher Education State Ministry of Education , Kaduna Kaduna State , Nigeria", "institution_ids": []}]}], "abstract_inverted_index": {"Recent": [0], "advances": [1], "in": [2, 21, 38, 143, 174], "long-text": [3], "language": [4, 23], "models": [5], "(LMs),": [6], "particularly": [7], "those": [8], "enhanced": [9], "with": [10, 51, 85, 132, 162], "agents": [11], "extending": [12], "Retrieval-Augmented": [13], "Generation": [14], "(RAG)": [15], "capabilities,": [16], "mark": [17], "a": [18, 41], "significant": [19], "leap": [20], "natural": [22], "processing": [24], "(NLP).": [25], "RAG": [26, 70, 157, 173], "enhances": [27], "traditional": [28], "LMs": [29], "by": [30], "retrieving": [31], "accurate": [32], "up-to-date": [33], "information": [34, 62], "from": [35], "external": [36], "sources": [37, 190], "real-time,": [39], "addressing": [40], "key": [42, 61], "limitation": [43], "of": [44, 170, 177], "Large": [45], "Language": [46], "Models": [47], "(LLMs)": [48], "that": [49], "struggle": [50], "broader": [52], "contextual": [53], "understanding.": [54], "Unlike": [55], "LLMs,": [56], "which": [57], "perform": [58], "best": [59], "when": [60], "is": [63], "at": [64], "the": [65, 175, 196], "input's": [66], "start": [67], "or": [68], "end,": [69], "enables": [71], "precise": [72], "responses": [73], "to": [74, 104, 156, 172, 184, 188], "complex,": [75], "context-dependent": [76], "queries.": [77], "This": [78, 181], "paper": [79, 160], "explores": [80], "RAG's": [81, 186], "advancements,": [82], "its": [83, 94, 102], "integration": [84], "technologies": [86], "like": [87], "virtual": [88], "and": [89, 93, 96, 108, 135, 147, 154, 191, 201], "augmented": [90], "reality": [91], "(VR/AR),": [92], "philosophical": [95], "ethical": [97], "implications.": [98], "We": [99, 149], "also": [100], "examine": [101], "potential": [103], "transform": [105], "education,": [106], "research,": [107], "professional": [109], "practice.": [110], "Additionally,": [111], "we": [112], "review": [113], "alternative": [114], "approaches,": [115], "including": [116], "Memory-Augmented": [117], "Neural": [118], "Networks": [119], "(MANNs),": [120], "knowledge-enhanced": [121], "LMs,": [122, 129], "dense": [123], "retrieval": [124], "models,": [125], "hybrid": [126], "systems,": [127], "long-context": [128], "generative": [130], "pretraining": [131], "prompt": [133], "engineering,": [134], "meta": [136], "learning.": [137], "These": [138], "alternatives": [139, 171], "offer": [140], "unique": [141], "advantages": [142], "efficiency,": [144], "scalability,": [145], "interpretability,": [146], "human-centricity.": [148], "compare": [150], "their": [151], "strengths,": [152], "limitations,": [153], "applications": [155], "systems.": [158, 204], "The": [159], "concludes": [161], "future": [163], "research": [164], "directions,": [165], "proposing": [166], "an": [167, 178], "optimized": [168], "ensemble": [169], "form": [176], "agency": [179], "construct.": [180], "approach": [182], "aims": [183], "extend": [185], "capabilities": [187], "dynamic": [189], "better-segmented": [192], "long": [193], "texts,": [194], "paving": [195], "way": [197], "for": [198], "more": [199], "robust": [200], "context-aware": [202], "AI": [203]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9810000061988831, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9810000061988831, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10286", "display_name": "Information Retrieval and Search Behavior", "score": 0.011300000362098217, "subfield": {"id": "https://openalex.org/subfields/1710", "display_name": "Information Systems"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.0032999999821186066, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.31219/osf.io/b94ng_v1", "is_oa": true, "landing_page_url": "https://doi.org/10.31219/osf.io/b94ng_v1", "pdf_url": "https://osf.io/b94ng_v1/download", "source": null, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4416371376", "doi": "https://doi.org/10.48550/arxiv.2503.10714", "title": "ZSMerge: Zero-Shot KV Cache Compression for Memory-Efficient Long-Context LLMs", "publication_year": 2025, "publication_date": "2025-03-13", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5090522322", "display_name": "Xin Liu", "orcid": "https://orcid.org/0000-0002-0270-375X", "observed_orcids": ["https://orcid.org/0000-0002-0270-375X"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Xin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5054245364", "display_name": "Xudong Wang", "orcid": "https://orcid.org/0000-0002-6575-5127", "observed_orcids": ["https://orcid.org/0000-0002-6575-5127"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wang, Xudong", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5029601647", "display_name": "Pei Liu", "orcid": "https://orcid.org/0000-0002-6854-4333", "observed_orcids": ["https://orcid.org/0000-0002-6854-4333"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Liu, Pei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5067691107", "display_name": "Guoming Tang", "orcid": "https://orcid.org/0000-0001-9801-1055", "observed_orcids": ["https://orcid.org/0000-0001-9801-1055"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Guoming", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0, 166], "linear": [1], "growth": [2], "of": [3, 146], "key-value": [4, 138], "(KV)": [5], "cache": [6, 29, 61, 67, 139], "memory": [7, 75, 116, 142], "and": [8, 99, 118], "quadratic": [9], "computational": [10], "in": [11, 23], "attention": [12, 97], "mechanisms": [13], "complexity": [14], "pose": [15], "significant": [16], "bottlenecks": [17], "for": [18, 65, 137], "large": [19], "language": [20], "models": [21], "(LLMs)": [22], "long-context": [24], "processing.": [25], "While": [26], "existing": [27], "KV": [28, 60], "optimization": [30], "methods": [31], "address": [32], "these": [33], "challenges": [34], "through": [35, 95], "token": [36, 80], "pruning": [37], "or": [38, 47], "feature": [39], "merging,": [40], "they": [41], "often": [42], "incur": [43], "irreversible": [44], "information": [45], "loss": [46], "require": [48], "costly": [49], "parameter": [50], "retraining.": [51, 112], "To": [52], "this": [53], "end,": [54], "we": [55], "propose": [56], "ZSMerge,": [57], "a": [58, 87, 101, 133], "dynamic": [59], "compression": [62, 135], "framework": [63], "designed": [64], "efficient": [66], "management,": [68], "featuring": [69], "three": [70], "key": [71], "operations:": [72], "(1)": [73], "fine-grained": [74], "allocation": [76], "guided": [77], "by": [78], "multi-dimensional": [79], "importance": [81], "metrics": [82], "at": [83, 158, 170], "head-level": [84], "granularity,": [85], "(2)": [86], "residual": [88], "merging": [89], "mechanism": [90, 104], "that": [91, 162], "preserves": [92], "critical": [93], "context": [94], "compensated": [96], "scoring,": [98], "(3)": [100], "zero-shot": [102], "adaptation": [103], "compatible": [105], "with": [106, 121, 154], "diverse": [107], "LLM": [108], "architectures": [109], "without": [110], "requiring": [111], "ZSMerge": [113], "significantly": [114], "enhances": [115], "efficiency": [117], "inference": [119], "speed": [120], "negligible": [122], "performance": [123], "degradation": [124], "across": [125], "LLMs.": [126], "When": [127], "applied": [128], "to": [129, 144], "LLaMA2-7B,": [130], "it": [131], "demonstrates": [132], "20:1": [134], "ratio": [136], "retention": [140], "(reducing": [141], "footprint": [143], "5\\%": [145], "baseline)": [147], "while": [148], "sustaining": [149], "comparable": [150], "generation": [151], "quality,": [152], "coupled": [153], "triple": [155], "throughput": [156], "gains": [157], "extreme": [159], "54k-token": [160], "contexts": [161], "eliminate": [163], "out-of-memory": [164], "failures.": [165], "code": [167], "is": [168], "available": [169], "https://github.com/SusCom-Lab/ZSMerge.": [171]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5971999764442444, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5971999764442444, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.29760000109672546, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.05090000107884407, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2503.10714", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2503.10714", "pdf_url": "https://arxiv.org/pdf/2503.10714", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4407683954", "doi": "https://doi.org/10.48550/arxiv.2502.10634", "title": "Lost in the Passage: Passage-level In-context Learning Does Not Necessarily Need a \"Passage\"", "publication_year": 2025, "publication_date": "2025-02-15", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5100556869", "display_name": "Hao Sun", "orcid": "https://orcid.org/0009-0001-1956-9921", "observed_orcids": ["https://orcid.org/0009-0001-1956-9921"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Sun, Hao", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5111019886", "display_name": "Chenming Tang", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Tang, Chenming", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5013025238", "display_name": "G.G. Li", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Li, Gengyang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5027803148", "display_name": "Yunfang Wu", "orcid": "https://orcid.org/0009-0001-9560-7512", "observed_orcids": ["https://orcid.org/0009-0001-9560-7512"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Wu, Yunfang", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"By": [0], "simply": [1, 150], "incorporating": [2], "demonstrations": [3], "into": [4], "the": [5, 40, 44, 48, 85, 108, 115, 121], "context,": [6], "in-context": [7], "learning": [8], "(ICL)": [9], "enables": [10], "large": [11], "language": [12], "models": [13], "(LLMs)": [14], "to": [15, 101, 104, 117], "yield": [16], "awesome": [17], "performance": [18, 83], "on": [19, 27, 57, 130, 139], "many": [20], "tasks.": [21], "In": [22], "this": [23], "study,": [24], "we": [25], "focus": [26], "passage-level": [28, 147], "long-context": [29, 141], "ICL": [30], "for": [31, 146], "generation": [32, 49, 60], "tasks": [33, 61, 142], "and": [34, 47, 66, 92, 110], "find": [35], "that": [36, 70, 96, 134], "LLMs": [37, 56, 97], "cannot": [38], "learn": [39], "intrinsic": [41], "relationship": [42], "between": [43], "demonstration": [45, 75, 154], "passage": [46, 76, 116], "output.": [50], "We": [51], "conduct": [52], "experiments": [53, 129], "with": [54, 77], "different": [55], "two": [58], "typical": [59], "including": [62], "single-document": [63], "question": [64], "answering": [65], "distractor": [67], "generation,": [68], "demonstrating": [69], "even": [71], "a": [72], "completely": [73], "meaningless": [74, 153], "1/4": [78], "length": [79], "achieves": [80, 157], "much": [81], "better": [82], "than": [84], "original": [86], "full": [87], "passage.": [88], "Analysis": [89], "via": [90], "attention": [91, 100], "information": [93, 112], "flow": [94], "reveals": [95], "pay": [98], "little": [99, 111], "passages": [102, 155], "compared": [103], "other": [105, 118, 140], "components": [106], "in": [107], "prompt": [109], "flows": [113], "from": [114], "parts": [119], "of": [120], "demonstration,": [122], "which": [123], "further": [124], "confirms": [125], "our": [126], "finding.": [127], "Additionally,": [128], "context": [131], "compression": [132, 135], "indicate": [133], "approaches": [136], "proven": [137], "effective": [138], "are": [143], "not": [144], "suitable": [145], "ICL,": [148], "since": [149], "using": [151], "shorter": [152], "already": [156], "competitive": [158], "performance.": [159]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9550999999046326, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.9550999999046326, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.0215000007301569, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.007499999832361937, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2502.10634", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2502.10634", "pdf_url": "https://arxiv.org/pdf/2502.10634", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4409030971", "doi": "https://doi.org/10.3389/frai.2025.1509338", "title": "Making sense of transformer success", "publication_year": 2025, "publication_date": "2025-04-01", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5007965265", "display_name": "Nicola Angius", "orcid": "https://orcid.org/0000-0002-7209-4400", "observed_orcids": ["https://orcid.org/0000-0002-7209-4400"]}, "institutions": [{"id": "https://openalex.org/I112862951", "display_name": "University of Messina", "ror": "https://ror.org/05ctdxz19", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I112862951"]}], "countries": ["IT"], "is_corresponding": false, "raw_author_name": "Nicola Angius", "raw_affiliation_strings": ["Department of Cognitive Science, University of Messina, Messina, Italy"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Cognitive Science, University of Messina, Messina, Italy", "institution_ids": ["https://openalex.org/I112862951"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5037233927", "display_name": "Pietro Perconti", "orcid": "https://orcid.org/0000-0002-3633-098X", "observed_orcids": ["https://orcid.org/0000-0002-3633-098X"]}, "institutions": [{"id": "https://openalex.org/I112862951", "display_name": "University of Messina", "ror": "https://ror.org/05ctdxz19", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I112862951"]}], "countries": ["IT"], "is_corresponding": false, "raw_author_name": "Pietro Perconti", "raw_affiliation_strings": ["Department of Cognitive Science, University of Messina, Messina, Italy"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Cognitive Science, University of Messina, Messina, Italy", "institution_ids": ["https://openalex.org/I112862951"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5020474912", "display_name": "Alessio Plebe", "orcid": "https://orcid.org/0000-0003-3666-061X", "observed_orcids": ["https://orcid.org/0000-0003-3666-061X"]}, "institutions": [{"id": "https://openalex.org/I112862951", "display_name": "University of Messina", "ror": "https://ror.org/05ctdxz19", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I112862951"]}], "countries": ["IT"], "is_corresponding": false, "raw_author_name": "Alessio Plebe", "raw_affiliation_strings": ["Department of Cognitive Science, University of Messina, Messina, Italy"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Cognitive Science, University of Messina, Messina, Italy", "institution_ids": ["https://openalex.org/I112862951"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5066466353", "display_name": "Alessandro Acciai", "orcid": "https://orcid.org/0000-0002-2547-9113", "observed_orcids": ["https://orcid.org/0000-0002-2547-9113"]}, "institutions": [{"id": "https://openalex.org/I112862951", "display_name": "University of Messina", "ror": "https://ror.org/05ctdxz19", "country_code": "IT", "type": "education", "lineage": ["https://openalex.org/I112862951"]}], "countries": ["IT"], "is_corresponding": true, "raw_author_name": "Alessandro Acciai", "raw_affiliation_strings": ["Department of Cognitive Science, University of Messina, Messina, Italy"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "Department of Cognitive Science, University of Messina, Messina, Italy", "institution_ids": ["https://openalex.org/I112862951"]}]}], "abstract_inverted_index": {"This": [0], "article": [1], "provides": [2], "an": [3, 74], "epistemological": [4], "analysis": [5, 141], "of": [6, 9, 17, 36, 38, 52, 86, 111, 123, 138, 145, 157, 167], "current": [7, 39, 171], "attempts": [8, 173], "explaining": [10], "how": [11, 45], "the": [12, 30, 34, 41, 46, 50, 84, 100, 121, 136, 139, 143, 148, 153, 200], "relatively": [13], "simple": [14], "algorithmic": [15], "components": [16], "neural": [18], "language": [19, 184], "models": [20], "(NLMs)": [21], "provide": [22, 163], "them": [23], "with": [24], "genuine": [25], "linguistic": [26], "competence.": [27], "After": [28], "introducing": [29], "Transformer": [31, 159], "architecture,": [32], "at": [33], "basis": [35], "most": [37], "NLMs,": [40], "paper": [42], "firstly": [43], "emphasizes": [44], "central": [47], "question": [48], "in": [49, 104, 131, 142, 195], "philosophy": [51, 144], "AI": [53], "has": [54], "been": [55], "shifted": [56], "from": [57, 99], "\u201ccan": [58], "machines": [59, 70], "think?\u201d,": [60, 71], "as": [61], "originally": [62], "put": [63], "by": [64], "Alan": [65], "Turing,": [66], "to": [67, 73, 90, 107, 119, 162, 174, 177, 188, 204], "\u201chow": [68], "can": [69], "pointing": [72], "explanatory": [75, 81, 101], "gap": [76], "for": [77, 83], "NLMs.": [78], "Subsequently,": [79], "existing": [80], "strategies": [82, 102], "functioning": [85], "NLMs": [87, 132, 176], "are": [88, 133, 160, 185, 202], "analyzed": [89], "argue": [91], "that": [92], "they,": [93], "however": [94], "debated,": [95], "do": [96], "not": [97], "differ": [98], "used": [103, 203], "cognitive": [105, 146], "science": [106], "explain": [108], "intelligent": [109], "behaviors": [110], "humans.": [112], "In": [113], "particular,": [114], "available": [115], "experimental": [116], "studies": [117], "turned": [118], "test": [120], "theory": [122], "mind,": [124], "discourse": [125], "entity": [126], "tracking,": [127], "and": [128, 152, 199, 206], "property": [129], "induction": [130, 154], "examined": [134], "under": [135], "light": [137], "functional": [140], "science;": [147], "so-called": [149], "copying": [150], "algorithm": [151], "head": [155], "phenomenon": [156], "a": [158, 164, 193, 197], "shown": [161, 187], "mechanist": [165], "explanation": [166], "in-context": [168], "learning;": [169], "finally,": [170], "pioneering": [172], "use": [175], "predict": [178], "brain": [179, 201], "activation": [180], "patterns": [181], "when": [182], "processing": [183], "here": [186], "involve": [189], "what": [190], "we": [191], "call": [192], "co-simulation,": [194], "which": [196], "NLM": [198], "simulate": [205], "understand": [207], "each": [208], "other.": [209]}, "primary_topic": {"id": "https://openalex.org/T10258", "display_name": "Philosophy and Theoretical Science", "score": 0.4291999936103821, "subfield": {"id": "https://openalex.org/subfields/3205", "display_name": "Experimental and Cognitive Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, "topics": [{"id": "https://openalex.org/T10258", "display_name": "Philosophy and Theoretical Science", "score": 0.4291999936103821, "subfield": {"id": "https://openalex.org/subfields/3205", "display_name": "Experimental and Cognitive Psychology"}, "field": {"id": "https://openalex.org/fields/32", "display_name": "Psychology"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T13308", "display_name": "Cybernetics and Technology in Society", "score": 0.3158999979496002, "subfield": {"id": "https://openalex.org/subfields/1207", "display_name": "History and Philosophy of Science"}, "field": {"id": "https://openalex.org/fields/12", "display_name": "Arts and Humanities"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T12026", "display_name": "Explainable Artificial Intelligence (XAI)", "score": 0.05590000003576279, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.3389/frai.2025.1509338", "is_oa": true, "landing_page_url": "https://doi.org/10.3389/frai.2025.1509338", "pdf_url": null, "source": {"id": "https://openalex.org/S4210197006", "display_name": "Frontiers in Artificial Intelligence", "issn_l": "2624-8212", "issn": ["2624-8212"], "is_oa": true, "is_in_doaj": true, "is_core": true, "listed_in": ["cwts-core", "doaj", "jufo-1", "ki-jl-1", "norway-1"], "host_organization": "https://openalex.org/P4310320527", "host_organization_name": "Frontiers Media", "host_organization_lineage": ["https://openalex.org/P4310320527"], "host_organization_lineage_names": ["Frontiers Media"], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Frontiers in Artificial Intelligence", "raw_type": "journal-article"}, "cited_by_count": 1, "is_retracted": false}, {"id": "https://openalex.org/W4411408314", "doi": "https://doi.org/10.1109/mgrs.2025.3574742", "title": "Enhancing Ultrahigh Resolution Remote Sensing Imagery Analysis With ImageRAG: A new framework", "publication_year": 2025, "publication_date": "2025-06-18", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5013896311", "display_name": "Zilun Zhang", "orcid": "https://orcid.org/0009-0008-5961-5970", "observed_orcids": ["https://orcid.org/0009-0008-5961-5970"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zilun Zhang", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": "https://orcid.org/0009-0008-5961-5970", "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Haozhan Shen", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Haozhan Shen", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5103955833", "display_name": "Tiancheng Zhao", "orcid": "https://orcid.org/0000-0002-4166-6189", "observed_orcids": ["https://orcid.org/0000-0002-4166-6189", "https://orcid.org/0009-0005-8726-2867"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Tiancheng Zhao", "raw_affiliation_strings": ["Binjiang Research Institute, Zhejiang University, Hangzhou, China"], "raw_orcid": "https://orcid.org/0000-0002-4166-6189", "affiliations": [{"raw_affiliation_string": "Binjiang Research Institute, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5110121539", "display_name": "Z. G. Guan", "orcid": "https://orcid.org/0009-0000-7335-4246", "observed_orcids": ["https://orcid.org/0009-0000-7335-4246"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Zian Guan", "raw_affiliation_strings": ["Polytechnic Institute, Zhejiang University, Hangzhou, China"], "raw_orcid": "https://orcid.org/0009-0000-7335-4246", "affiliations": [{"raw_affiliation_string": "Polytechnic Institute, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5092494960", "display_name": "Bin Chen", "orcid": "https://orcid.org/0000-0002-8028-0182", "observed_orcids": ["https://orcid.org/0000-0002-8028-0182"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Bin Chen", "raw_affiliation_strings": ["School of Software Engineering, Zhejiang University, Hangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Software Engineering, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Yuhao Wang", "orcid": "https://orcid.org/0009-0000-6753-5008", "observed_orcids": ["https://orcid.org/0009-0000-6753-5008"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yuhao Wang", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": "https://orcid.org/0009-0000-6753-5008", "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5042055991", "display_name": "Jia Wei Xu", "orcid": "https://orcid.org/0000-0003-4061-8262", "observed_orcids": ["https://orcid.org/0000-0003-4061-8262", "https://orcid.org/0009-0007-1298-8086"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Xu Jia", "raw_affiliation_strings": ["School of Software Engineering, Zhejiang University, Hangzhou, China"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "School of Software Engineering, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5040589965", "display_name": "Yuxiang Cai", "orcid": "https://orcid.org/0000-0003-2784-6607", "observed_orcids": ["https://orcid.org/0000-0003-2784-6607"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yuxiang Cai", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": "https://orcid.org/0000-0003-2784-6607", "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5010234425", "display_name": "Yongheng Shang", "orcid": "https://orcid.org/0000-0001-8550-6710", "observed_orcids": ["https://orcid.org/0000-0001-8550-6710"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Yongheng Shang", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": "https://orcid.org/0000-0001-8550-6710", "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5069353502", "display_name": "Jianwei Yin", "orcid": "https://orcid.org/0000-0003-4703-7348", "observed_orcids": ["https://orcid.org/0000-0003-4703-7348"]}, "institutions": [{"id": "https://openalex.org/I76130692", "display_name": "Zhejiang University", "ror": "https://ror.org/00a2xv884", "country_code": "CN", "type": "education", "lineage": ["https://openalex.org/I76130692"]}], "countries": ["CN"], "is_corresponding": false, "raw_author_name": "Jianwei Yin", "raw_affiliation_strings": ["College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "College of Computer Science and TechnologyZhejiang University"], "raw_orcid": "https://orcid.org/0000-0003-4703-7348", "affiliations": [{"raw_affiliation_string": "College of Computer Science and Technology, Zhejiang University, Hangzhou, China", "institution_ids": ["https://openalex.org/I76130692"]}, {"raw_affiliation_string": "College of Computer Science and TechnologyZhejiang University", "institution_ids": ["https://openalex.org/I76130692"]}]}], "abstract_inverted_index": {"Ultrahigh": [0], "resolution": [1], "(UHR)": [2], "remote": [3], "sensing": [4], "imagery": [5], "(RSI)": [6], "(e.g.": [7], "10,000": [8, 10], "\u00d7": [9], "pixels)": [11], "poses": [12], "a": [13, 94, 105, 168], "significant": [14], "challenge": [15], "for": [16, 92], "current": [17], "RS": [18, 113], "vision-language": [19], "models": [20], "(RSVLMs).": [21], "If": [22], "one": [23], "chooses": [24], "to": [25, 30, 67, 76, 96, 117, 148, 167, 182, 192], "resize": [26], "the": [27, 31, 36, 50, 58, 69, 78, 82, 98, 111, 118, 133, 154, 159, 204], "UHR": [28, 43, 102, 112, 160, 200], "image": [29, 34, 71, 114, 127, 161], "standard": [32, 62], "input": [33], "size,": [35], "extensive": [37, 194], "spatial": [38, 197], "and": [39, 72, 151, 173, 187, 196, 209], "contextual": [40, 128], "information": [41, 198], "that": [42, 165, 203], "images": [44, 55], "contain": [45], "will": [46, 213], "be": [47, 214], "neglected.": [48], "Otherwise,": [49], "original": [51], "size": [52], "of": [53, 61, 100, 158], "these": [54], "often": [56], "exceeds": [57], "token": [59], "limits": [60], "RSVLMs,": [63], "making": [64], "it": [65], "difficult": [66], "process": [68], "entire": [70], "capture": [73], "long-range": [74], "dependencies": [75], "answer": [77], "query": [79], "based": [80, 131], "on": [81, 132, 153], "abundant": [83], "visual": [84, 163], "context.": [85], "In": [86], "this": [87, 180, 184], "article,": [88], "we": [89, 123], "introduce": [90], "ImageRAG": [91, 189], "RS,": [93], "framework": [95, 181], "address": [97], "complexities": [99], "analyzing": [101], "RSI": [103], "with": [104], "little": [106], "training": [107], "requirement.": [108], "By": [109], "transforming": [110], "analysis": [115, 205], "task": [116, 185], "image\u2019s": [119], "long-context": [120], "selection": [121], "task,": [122], "design": [124], "an": [125], "innovative": [126], "retrieval": [129], "mechanism": [130], "retrieval-augmented": [134], "generation": [135], "(RAG)": [136], "technique,": [137], "denoted": [138], "as": [139, 162], "ImageRAG.": [140], "ImageRAG\u2019s": [141], "core": [142], "innovation": [143], "lies": [144], "in": [145, 179], "its": [146], "ability": [147], "selectively": [149], "retrieve": [150], "focus": [152], "most": [155], "relevant": [156], "portions": [157], "contexts": [164], "pertain": [166], "given": [169], "query.": [170], "Fast": [171], "path": [172, 175], "slow": [174], "modes": [176], "are": [177], "proposed": [178], "handle": [183], "efficiently": [186], "effectively.": [188], "allows": [190], "RSVLMs": [191], "manage": [193], "context": [195], "from": [199], "RSI,": [201], "ensuring": [202], "is": [206], "both": [207], "accurate": [208], "efficient.": [210], "The": [211], "code": [212], "released": [215], "athttps://github.com/om-ai-lab/ImageRAG.": [216]}, "primary_topic": {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.5503000020980835, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.5503000020980835, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10689", "display_name": "Remote-Sensing Image Classification", "score": 0.2451999932527542, "subfield": {"id": "https://openalex.org/subfields/2214", "display_name": "Media Technology"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.04479999840259552, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.1109/mgrs.2025.3574742", "is_oa": false, "landing_page_url": "https://doi.org/10.1109/mgrs.2025.3574742", "pdf_url": null, "source": {"id": "https://openalex.org/S2491948244", "display_name": "IEEE Geoscience and Remote Sensing Magazine", "issn_l": "2168-6831", "issn": ["2168-6831", "2373-7468", "2473-2397"], "is_oa": false, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "jufo-2"], "host_organization": "https://openalex.org/P4310319808", "host_organization_name": "Institute of Electrical and Electronics Engineers", "host_organization_lineage": ["https://openalex.org/P4310319808"], "host_organization_lineage_names": ["Institute of Electrical and Electronics Engineers"], "type": "journal"}, "license": null, "license_id": null, "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "IEEE Geoscience and Remote Sensing Magazine", "raw_type": "journal-article"}, "cited_by_count": 7, "is_retracted": false}, {"id": "https://openalex.org/W7106699545", "doi": "https://doi.org/10.48550/arxiv.2511.17666", "title": "Evaluating Adversarial Vulnerabilities in Modern Large Language Models", "publication_year": 2025, "publication_date": "2025-11-21", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Perel, Tom", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": true, "raw_author_name": "Perel, Tom", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"The": [0, 62, 126, 146], "recent": [1], "boom": [2], "and": [3, 24, 49, 80, 105, 124, 157, 196], "rapid": [4], "integration": [5], "of": [6, 15, 21, 33, 113, 128, 136, 178, 204, 211], "Large": [7], "Language": [8], "Models": [9], "(LLMs)": [10], "into": [11, 200], "a": [12, 18, 30, 143, 149, 189], "wide": [13], "range": [14], "applications": [16], "warrants": [17], "deeper": [19], "understanding": [20], "their": [22, 76, 162], "security": [23], "safety": [25, 78, 163, 217], "vulnerabilities.": [26], "This": [27, 186], "paper": [28], "presents": [29], "comparative": [31], "analysis": [32], "the": [34, 53, 59, 92, 129, 134, 182, 201, 208], "susceptibility": [35, 153], "to": [36, 74, 88, 108], "jailbreak": [37, 152], "attacks": [38, 169], "for": [39, 192], "two": [40, 65], "leading": [41], "publicly": [42], "available": [43], "LLMs,": [44], "Google's": [45], "Gemini": [46], "2.5": [47, 155], "Flash": [48, 156], "OpenAI's": [50], "GPT-4": [51], "(specifically": [52], "GPT-4o": [54], "mini": [55], "model": [56, 84, 213], "accessible": [57], "in": [58, 91, 151, 161, 181], "free": [60], "tier).": [61], "research": [63, 187], "utilized": [64], "main": [66], "bypass": [67], "strategies:": [68], "'self-bypass',": [69], "where": [70, 82], "models": [71], "were": [72, 97, 170], "prompted": [73], "circumvent": [75], "own": [77], "protocols,": [79], "'cross-bypass',": [81], "one": [83], "generated": [85], "adversarial": [86], "prompts": [87], "exploit": [89], "vulnerabilities": [90, 179], "other.": [93], "Four": [94], "attack": [95, 130], "methods": [96], "employed": [98], "-": [99, 107], "direct": [100], "injection,": [101], "role-playing,": [102], "context": [103], "manipulation,": [104], "obfuscation": [106], "generate": [109], "five": [110], "distinct": [111], "categories": [112], "unsafe": [114], "content:": [115], "hate": [116], "speech,": [117], "illegal": [118], "activities,": [119], "malicious": [120], "code,": [121], "dangerous": [122], "content,": [123, 138], "misinformation.": [125], "success": [127], "was": [131], "determined": [132], "by": [133], "generation": [135], "disallowed": [137], "with": [139, 215], "successful": [140], "jailbreaks": [141], "assigned": [142], "severity": [144], "score.": [145], "findings": [147], "indicate": [148], "disparity": [150], "between": [154], "GPT-4,": [158], "suggesting": [159], "variations": [160], "implementations": [164], "or": [165], "architectural": [166], "design.": [167], "Cross-bypass": [168], "particularly": [171], "effective,": [172], "indicating": [173], "that": [174], "an": [175], "ample": [176], "amount": [177], "exist": [180], "underlying": [183], "transformer": [184], "architecture.": [185], "contributes": [188], "scalable": [190], "framework": [191], "automated": [193], "AI": [194], "red-teaming": [195], "provides": [197], "data-driven": [198], "insights": [199], "current": [202], "state": [203], "LLM": [205], "safety,": [206], "underscoring": [207], "complex": [209], "challenge": [210], "balancing": [212], "capabilities": [214], "robust": [216], "mechanisms.": [218]}, "primary_topic": {"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 1.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T12262", "display_name": "Hate Speech and Cyberbullying Detection", "score": 0.0, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2511.17666", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2511.17666", "pdf_url": "https://arxiv.org/pdf/2511.17666", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415491770", "doi": "https://doi.org/10.36227/techrxiv.176127449.98578610/v1", "title": "Green Transformers: Leveraging Neuromorphic Computing, Sparse Attention, and Adaptive Quantization for Energy-Efficient Inference", "publication_year": 2025, "publication_date": "2025-10-24", "type": "preprint", "language": null, "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5120115105", "display_name": "Ankit Kumar Mandusia", "orcid": "https://orcid.org/0009-0008-0005-083X", "observed_orcids": ["https://orcid.org/0009-0008-0005-083X"]}, "institutions": [{"id": "https://openalex.org/I178254495", "display_name": "Dayalbagh Educational Institute", "ror": "https://ror.org/04q4j2f69", "country_code": "IN", "type": "education", "lineage": ["https://openalex.org/I178254495"]}], "countries": ["IN"], "is_corresponding": false, "raw_author_name": "Ankit Kumar Mandusia", "raw_affiliation_strings": ["Department of Physics and Computer Science , Dayalbagh Educational Institute , Agra , India"], "raw_orcid": "https://orcid.org/0009-0008-0005-083X", "affiliations": [{"raw_affiliation_string": "Department of Physics and Computer Science , Dayalbagh Educational Institute , Agra , India", "institution_ids": ["https://openalex.org/I178254495"]}]}, {"author_position": "last", "author": {"id": null, "display_name": "Bhanu Pratap Rana", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Bhanu Pratap Rana", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Large": [0], "language": [1, 8], "models": [2, 32], "achieve": [3], "remarkable": [4], "performance": [5], "in": [6, 89], "natural": [7], "processing": [9], "but": [10], "consume": [11], "significant": [12], "energy,": [13], "with": [14, 48, 126], "inference": [15, 82, 121], "accounting": [16], "for": [17, 31, 110, 119], "60-90%": [18], "of": [19, 28], "their": [20], "environmental": [21], "footprint,": [22], "emitting": [23], "up": [24, 74], "to": [25, 51, 70, 75, 92], "6,930": [26], "tons": [27], "CO2": [29], "equivalent": [30], "like": [33, 129], "LLaMA": [34], "3.": [35], "This": [36], "paper": [37], "introduces": [38], "a": [39, 115], "Hybrid": [40], "Neuromorphic": [41], "Transformer": [42, 49], "that": [43], "integrates": [44], "spiking": [45], "neural": [46], "networks": [47], "architectures": [50], "enhance": [52], "energy": [53, 77], "efficiency.": [54], "A": [55], "Spiking": [56], "Self-Attention": [57], "module": [58], "reduces": [59], "computational": [60], "complexity,": [61, 72], "while": [62], "an": [63], "Adaptive": [64], "Quantization": [65], "Scheduler": [66], "dynamically": [67], "adjusts": [68], "precision": [69], "input": [71], "achieving": [73], "59%": [76], "savings.": [78], "Carbon-Aware": [79], "Scheduling": [80], "optimizes": [81], "across": [83], "data": [84], "centers,": [85], "leveraging": [86], "geographic": [87], "variations": [88], "carbon": [90], "intensity": [91], "cut": [93], "emissions": [94], "by": [95], "13%.": [96], "Evaluated": [97], "on": [98, 112, 122], "GLUE": [99], "tasks,": [100], "our": [101], "approach": [102], "maintains": [103], "high": [104], "accuracy,": [105], "such": [106], "as": [107], "0.9036": [108], "F1": [109], "BERT": [111], "SST-2,": [113], "offering": [114], "scalable,": [116], "sustainable": [117], "framework": [118], "neuromorphiccompatible": [120], "resource-constrained": [123], "devices,": [124], "aligning": [125], "emerging": [127], "regulations": [128], "the": [130], "EU": [131], "AI": [132], "Act.": [133]}, "primary_topic": {"id": "https://openalex.org/T12238", "display_name": "Green IT and Sustainability", "score": 0.4027999937534332, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T12238", "display_name": "Green IT and Sustainability", "score": 0.4027999937534332, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10502", "display_name": "Advanced Memory and Neural Computing", "score": 0.35280001163482666, "subfield": {"id": "https://openalex.org/subfields/2208", "display_name": "Electrical and Electronic Engineering"}, "field": {"id": "https://openalex.org/fields/22", "display_name": "Engineering"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.09700000286102295, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.36227/techrxiv.176127449.98578610/v1", "is_oa": true, "landing_page_url": "https://doi.org/10.36227/techrxiv.176127449.98578610/v1", "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.176127449.98578610/v1", "source": null, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": null, "raw_type": "posted-content"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4415401063", "doi": "https://doi.org/10.3390/electronics14204122", "title": "Catch Me If You Can: Rogue AI Detection and Correction at Scale", "publication_year": 2025, "publication_date": "2025-10-21", "type": "article", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5009120290", "display_name": "Fatemeh Stodt", "orcid": "https://orcid.org/0000-0003-0863-0907", "observed_orcids": ["https://orcid.org/0000-0003-0863-0907"]}, "institutions": [{"id": "https://openalex.org/I882953328", "display_name": "Furtwangen University", "ror": "https://ror.org/02m11x738", "country_code": "DE", "type": "education", "lineage": ["https://openalex.org/I882953328"]}], "countries": ["DE"], "is_corresponding": true, "raw_author_name": "Fatemeh Stodt", "raw_affiliation_strings": ["Institute for Data Science, Cloud Computing and IT Security (IDACUS), University of Applied Sciences Furtwangen, 78120 Furtwangen, Germany"], "raw_orcid": "https://orcid.org/0000-0003-0863-0907", "affiliations": [{"raw_affiliation_string": "Institute for Data Science, Cloud Computing and IT Security (IDACUS), University of Applied Sciences Furtwangen, 78120 Furtwangen, Germany", "institution_ids": ["https://openalex.org/I882953328"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5066802834", "display_name": "Jan Stodt", "orcid": "https://orcid.org/0000-0001-9115-7668", "observed_orcids": ["https://orcid.org/0000-0001-9115-7668"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Jan Stodt", "raw_affiliation_strings": ["go AVA GmbH, 45136 Essen, Germany"], "raw_orcid": "https://orcid.org/0000-0001-9115-7668", "affiliations": [{"raw_affiliation_string": "go AVA GmbH, 45136 Essen, Germany", "institution_ids": []}]}, {"author_position": "middle", "author": {"id": null, "display_name": "Mohammed Alshawki", "orcid": "https://orcid.org/0000-0003-1619-2927", "observed_orcids": ["https://orcid.org/0000-0003-1619-2927"]}, "institutions": [{"id": "https://openalex.org/I106118109", "display_name": "E\u00f6tv\u00f6s Lor\u00e1nd University", "ror": "https://ror.org/01jsq2704", "country_code": "HU", "type": "education", "lineage": ["https://openalex.org/I106118109"]}], "countries": ["HU"], "is_corresponding": false, "raw_author_name": "Mohammed Alshawki", "raw_affiliation_strings": ["Department of Computer Algebra, E\u00f6tv\u00f6s Lor\u00e1nd University, 1053 Budapest, Hungary"], "raw_orcid": "https://orcid.org/0000-0003-1619-2927", "affiliations": [{"raw_affiliation_string": "Department of Computer Algebra, E\u00f6tv\u00f6s Lor\u00e1nd University, 1053 Budapest, Hungary", "institution_ids": ["https://openalex.org/I106118109"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5057946130", "display_name": "Javad Salimi Sartakhti", "orcid": "https://orcid.org/0000-0003-1183-1232", "observed_orcids": ["https://orcid.org/0000-0003-1183-1232"]}, "institutions": [{"id": "https://openalex.org/I105818395", "display_name": "University of Kashan", "ror": "https://ror.org/015zmr509", "country_code": "IR", "type": "education", "lineage": ["https://openalex.org/I105818395"]}], "countries": ["IR"], "is_corresponding": false, "raw_author_name": "Javad Salimi Sratakhti", "raw_affiliation_strings": ["Department of Electrical and Computer Engineering, University of Kashan, Kashan 87317, Iran"], "raw_orcid": "https://orcid.org/0000-0003-1183-1232", "affiliations": [{"raw_affiliation_string": "Department of Electrical and Computer Engineering, University of Kashan, Kashan 87317, Iran", "institution_ids": ["https://openalex.org/I105818395"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5088114580", "display_name": "Christoph Reich", "orcid": "https://orcid.org/0000-0001-9831-2181", "observed_orcids": ["https://orcid.org/0000-0001-9831-2181"]}, "institutions": [{"id": "https://openalex.org/I882953328", "display_name": "Furtwangen University", "ror": "https://ror.org/02m11x738", "country_code": "DE", "type": "education", "lineage": ["https://openalex.org/I882953328"]}], "countries": ["DE"], "is_corresponding": false, "raw_author_name": "Christoph Reich", "raw_affiliation_strings": ["Institute for Data Science, Cloud Computing and IT Security (IDACUS), University of Applied Sciences Furtwangen, 78120 Furtwangen, Germany"], "raw_orcid": "https://orcid.org/0000-0001-9831-2181", "affiliations": [{"raw_affiliation_string": "Institute for Data Science, Cloud Computing and IT Security (IDACUS), University of Applied Sciences Furtwangen, 78120 Furtwangen, Germany", "institution_ids": ["https://openalex.org/I882953328"]}]}], "abstract_inverted_index": {"Modern": [0], "AI": [1], "systems": [2], "can": [3, 234], "strategically": [4], "misreport": [5, 260], "information": [6], "when": [7], "incentives": [8], "diverge": [9], "from": [10, 266], "truthfulness,": [11], "posing": [12], "risks": [13], "for": [14, 49, 215, 253, 277], "oversight": [15], "and": [16, 66, 83, 96, 144, 196, 212, 243, 247, 262], "deployment.": [17, 278], "Prior": [18], "studies": [19], "often": [20], "examine": [21], "this": [22, 179], "behavior": [23], "within": [24], "a": [25, 32, 45, 84, 101, 106, 125, 156, 240, 269], "single": [26], "paradigm;": [27], "systematic,": [28], "cross-architecture": [29, 127], "evidence": [30], "under": [31, 199, 239], "unified": [33], "protocol": [34], "has": [35], "been": [36], "limited.": [37], "We": [38, 255], "introduce": [39], "the": [40, 112, 119, 181], "Strategy": [41], "Elicitation": [42], "Battery": [43], "(SEB),": [44], "standardized": [46], "probe": [47, 85], "suite": [48], "measuring": [50], "deceptive": [51], "reporting": [52, 245], "across": [53], "large": [54], "language": [55], "models": [56], "(LLMs),": [57], "reinforcement-learning": [58], "agents,": [59], "vision-only": [60], "classifiers,": [61], "multimodal": [62], "encoders,": [63], "state-space": [64], "models,": [65], "diffusion": [67], "models.": [68], "SEB": [69, 123], "uses": [70], "Bayesian": [71], "inference": [72], "tasks": [73], "with": [74, 100], "persona-controlled": [75], "instructions,": [76], "schema-constrained": [77], "outputs,": [78], "deterministic": [79], "decoding": [80], "where": [81], "supported,": [82], "mix": [86], "(near-threshold,": [87], "repeats,": [88], "neutralized,": [89], "cross-checks).": [90], "Estimates": [91], "use": [92], "clustered": [93], "bootstrap": [94], "intervals,": [95], "significance": [97], "is": [98, 109, 171], "assessed": [99], "logistic": [102, 150], "regression": [103, 151], "by": [104, 203], "architecture;": [105], "mixed-effects": [107], "analysis": [108], "planned": [110], "once": [111], "per-round": [113], "agent/episode": [114], "traces": [115], "are": [116, 184, 190, 224, 250], "exported.": [117], "On": [118], "latest": [120], "pre-correction": [121], "runs,": [122], "shows": [124], "consistent": [126], "pattern": [128], "in": [129, 178], "deception": [130, 195, 222], "rates:": [131], "ViT": [132, 170], "80.0%,": [133], "CLIP": [134], "15.0%,": [135], "Mamba": [136], "10.0%,": [137, 140, 143], "RL": [138], "agents": [139], "Stable": [141], "Diffusion": [142], "LLMs": [145], "5.0%": [146], "(20": [147], "scenarios/architecture).": [148], "A": [149], "on": [152], "per-scenario": [153], "flags": [154, 223], "finds": [155], "significant": [157], "overall": [158], "architecture": [159, 204], "effect": [160], "(likelihood-ratio": [161], "test": [162], "vs.": [163], "intercept-only:": [164], "\u03c72(5)=41.56,": [165], "p=7.22\u00d710\u22128).": [166], "Holm-adjusted": [167], "contrasts": [168], "indicate": [169, 219], "significantly": [172], "higher": [173], "than": [174], "all": [175], "other": [176], "architectures": [177, 233], "snapshot;": [180], "remaining": [182], "pairs": [183], "not": [185, 225], "significant.": [186], "Post-correction": [187], "acceptance": [188, 275], "decisions": [189, 276], "evaluated": [191], "separately": [192], "using": [193], "residual": [194], "override": [197], "rates": [198], "SEB-Correct.": [200], "Latency": [201], "varies": [202], "(sub-second": [205], "to": [206, 227, 236], "minutes),": [207], "enabling": [208], "pre-deployment": [209], "screening": [210], "broadly": [211], "real-time": [213], "auditing": [214], "low-latency": [216], "classes.": [217], "Results": [218], "that": [220, 231, 244], "SEB-Detect": [221], "confined": [226], "any": [228], "one": [229], "paradigm,": [230], "distinct": [232], "converge": [235], "similar": [237], "levels": [238], "common": [241], "interface,": [242], "interfaces": [246], "incentive": [248], "framing": [249], "central": [251], "levers": [252], "mitigation.": [254], "operationalize": [256], "\u201cdeception\u201d": [257], "as": [258], "reward-sensitive": [259], "flags,": [261], "we": [263], "separate": [264], "detection": [265], "intervention": [267], "via": [268], "correction": [270], "wrapper": [271], "(SEB-Correct),": [272], "supporting": [273], "principled": [274]}, "primary_topic": {"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.3871999979019165, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11689", "display_name": "Adversarial Robustness in Machine Learning", "score": 0.3871999979019165, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10883", "display_name": "Ethics and Social Impacts of AI", "score": 0.28780001401901245, "subfield": {"id": "https://openalex.org/subfields/3311", "display_name": "Safety Research"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}, {"id": "https://openalex.org/T11147", "display_name": "Misinformation and Its Impacts", "score": 0.22769999504089355, "subfield": {"id": "https://openalex.org/subfields/3312", "display_name": "Sociology and Political Science"}, "field": {"id": "https://openalex.org/fields/33", "display_name": "Social Sciences"}, "domain": {"id": "https://openalex.org/domains/2", "display_name": "Social Sciences"}}], "primary_location": {"id": "doi:10.3390/electronics14204122", "is_oa": true, "landing_page_url": "https://doi.org/10.3390/electronics14204122", "pdf_url": "https://www.mdpi.com/2079-9292/14/20/4122/pdf?version=1761049320", "source": {"id": "https://openalex.org/S4210202905", "display_name": "Electronics", "issn_l": "2079-9292", "issn": ["2079-9292"], "is_oa": true, "is_in_doaj": false, "is_core": true, "listed_in": ["cwts-core", "norway-1"], "host_organization": "https://openalex.org/P4310310987", "host_organization_name": "Multidisciplinary Digital Publishing Institute", "host_organization_lineage": ["https://openalex.org/P4310310987"], "host_organization_lineage_names": ["Multidisciplinary Digital Publishing Institute"], "type": "journal"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": "publishedVersion", "is_accepted": true, "is_published": true, "raw_source_name": "Electronics", "raw_type": "journal-article"}, "cited_by_count": 2, "is_retracted": false}, {"id": "https://openalex.org/W7108811324", "doi": "https://doi.org/10.5281/zenodo.17815797", "title": "Semantic-Aware Caching for Memory-Efficient Large Language Model Inference", "publication_year": 2025, "publication_date": "2025-12-04", "type": "article", "language": null, "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Revista, Zen", "orcid": "https://orcid.org/0009-0007-6299-2008", "observed_orcids": ["https://orcid.org/0009-0007-6299-2008"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Revista, Zen", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0007-6299-2008", "affiliations": []}, {"author_position": "last", "author": {"id": null, "display_name": "IA, 10", "orcid": "https://orcid.org/0009-0007-6299-2008", "observed_orcids": ["https://orcid.org/0009-0007-6299-2008"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "IA, 10", "raw_affiliation_strings": [], "raw_orcid": "https://orcid.org/0009-0007-6299-2008", "affiliations": []}], "abstract_inverted_index": {"Large": [0], "Language": [1], "Models": [2], "(LLMs)": [3], "have": [4], "demonstrated": [5], "remarkable": [6], "capabilities": [7], "across": [8], "various": [9], "natural": [10], "language": [11], "processing": [12], "tasks,": [13], "yet": [14], "their": [15, 117, 141], "deployment": [16, 227], "in": [17, 143, 231], "real-world": [18], "applications": [19, 236], "is": [20, 39], "often": [21], "constrained": [22], "by": [23], "significant": [24], "memory": [25, 37, 91, 193], "requirements,": [26], "particularly": [27], "during": [28, 57], "the": [29, 40, 64, 80, 110, 146, 150, 216], "inference": [30, 71, 203], "phase.": [31], "A": [32], "major": [33], "contributor": [34], "to": [35, 53, 89, 123, 191, 221], "this": [36], "bottleneck": [38], "Key-Value": [41], "(KV)": [42], "cache,": [43], "which": [44], "stores": [45], "intermediate": [46], "attention": [47], "states": [48], "for": [49, 93, 159, 164, 235], "previously": [50], "processed": [51], "tokens": [52], "avoid": [54], "redundant": [55], "computations": [56], "autoregressive": [58], "decoding.": [59], "As": [60], "context": [61, 239], "windows": [62], "expand,": [63], "KV": [65, 113, 129], "cache": [66, 130], "grows": [67], "linearly,": [68], "making": [69], "long-context": [70], "a": [72, 86, 157, 208], "memory-bound": [73], "problem.": [74], "This": [75], "paper": [76, 206], "proposes": [77], "and": [78, 121, 137, 174, 202, 211, 225], "explores": [79], "concept": [81], "of": [82, 112, 149, 218, 228], "semantic-aware": [83, 107, 160, 219], "caching": [84, 98, 108, 220], "as": [85], "novel": [87], "approach": [88, 188], "improve": [90], "efficiency": [92], "LLM": [94], "inference.": [95], "Unlike": [96], "traditional": [97], "strategies": [99], "that": [100], "rely": [101], "on": [102, 116, 180], "recency": [103], "or": [104, 197, 234], "fixed-size": [105], "policies,": [106], "prioritizes": [109], "retention": [111], "pairs": [114], "based": [115, 179], "perceived": [118], "semantic": [119, 147, 181], "importance": [120], "relevance": [122], "future": [124], "generations.": [125], "We": [126, 183], "review": [127], "existing": [128], "optimization": [131], "techniques,": [132], "including": [133], "quantization,": [134], "eviction": [135], "strategies,": [136], "architectural": [138], "innovations,": [139], "highlighting": [140], "limitations": [142], "fully": [144], "leveraging": [145], "content": [148], "cached": [151], "information.": [152], "Our": [153], "proposed": [154], "methodology": [155], "outlines": [156], "framework": [158, 210], "caching,": [161], "detailing": [162], "mechanisms": [163], "dynamically": [165], "scoring": [166], "token": [167], "importance,": [168], "adaptively": [169], "evicting": [170], "less": [171], "critical": [172], "tokens,": [173], "potentially": [175], "applying": [176], "heterogeneous": [177], "compression": [178], "value.": [182], "discuss": [184], "how": [185], "such": [186], "an": [187], "can": [189], "lead": [190], "substantial": [192], "reductions": [194], "while": [195], "preserving": [196], "even": [198], "enhancing": [199], "model": [200], "quality": [201], "throughput.": [204], "The": [205], "presents": [207], "conceptual": [209], "discusses": [212], "expected": [213], "results,": [214], "demonstrating": [215], "potential": [217], "enable": [222], "more": [223], "efficient": [224], "scalable": [226], "LLMs,": [229], "especially": [230], "resource-constrained": [232], "environments": [233], "requiring": [237], "extensive": [238], "understanding.": [240]}, "primary_topic": {"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5503000020980835, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T10028", "display_name": "Topic Modeling", "score": 0.5503000020980835, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10181", "display_name": "Natural Language Processing Techniques", "score": 0.2671000063419342, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 0.07739999890327454, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "doi:10.5281/zenodo.17815797", "is_oa": true, "landing_page_url": "https://doi.org/10.5281/zenodo.17815797", "pdf_url": null, "source": {"id": "https://openalex.org/S4306400562", "display_name": "Zenodo (CERN European Organization for Nuclear Research)", "issn_l": null, "issn": null, "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I67311998", "host_organization_name": "European Organization for Nuclear Research", "host_organization_lineage": ["https://openalex.org/I67311998"], "host_organization_lineage_names": [], "type": "repository"}, "license": "cc-by", "license_id": "https://openalex.org/licenses/cc-by", "version": null, "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "JournalArticle"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4417521603", "doi": "https://doi.org/10.48550/arxiv.2504.21063", "title": "Token-Level Prompt Mixture with Parameter-Free Routing for Federated Domain Generalization", "publication_year": 2025, "publication_date": "2025-04-29", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": null, "display_name": "Gong, Shuai", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Gong, Shuai", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5040285913", "display_name": "Chaoran Cui", "orcid": "https://orcid.org/0000-0003-3332-1348", "observed_orcids": ["https://orcid.org/0000-0003-3332-1348"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Cui, Chaoran", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": null, "display_name": "Dong, Xiaolin", "orcid": null, "observed_orcids": []}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Dong, Xiaolin", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5019824207", "display_name": "Xiushan Nie", "orcid": "https://orcid.org/0000-0001-9644-9723", "observed_orcids": ["https://orcid.org/0000-0001-9644-9723"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Nie, Xiushan", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5108048954", "display_name": "Lei Zhu", "orcid": "https://orcid.org/0000-0002-2993-7142", "observed_orcids": ["https://orcid.org/0000-0002-2993-7142"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Zhu, Lei", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}, {"author_position": "last", "author": {"id": "https://openalex.org/A5034967388", "display_name": "Xiaojun Chang", "orcid": "https://orcid.org/0000-0002-7778-8807", "observed_orcids": ["https://orcid.org/0000-0002-7778-8807"]}, "institutions": [], "countries": [], "is_corresponding": false, "raw_author_name": "Chang, Xiaojun", "raw_affiliation_strings": [], "raw_orcid": null, "affiliations": []}], "abstract_inverted_index": {"Federated": [0], "domain": [1], "generalization": [2, 169, 181], "(FedDG)": [3], "aims": [4], "to": [5, 26, 47, 117, 153], "learn": [6], "a": [7, 35, 41, 60, 88, 126], "globally": [8], "generalizable": [9], "model": [10], "from": [11, 69, 78], "decentralized": [12], "clients": [13], "with": [14, 92, 183], "heterogeneous": [15], "data": [16], "while": [17], "preserving": [18], "privacy.": [19], "Recent": [20], "studies": [21], "have": [22], "introduced": [23], "prompt": [24, 90, 139, 163], "learning": [25, 34, 43, 160], "adapt": [27], "vision-language": [28], "models": [29], "(VLMs)": [30], "in": [31], "FedDG": [32], "by": [33, 143, 147], "single": [36], "global": [37], "prompt.": [38], "However,": [39], "such": [40], "one-prompt-fits-all": [42], "paradigm": [44], "typically": [45], "leads": [46], "performance": [48], "degradation": [49], "on": [50, 131], "personalized": [51], "samples.": [52], "Although": [53], "the": [54, 148, 166], "mixture": [55, 91], "of": [56, 150, 185], "experts": [57], "(MoE)": [58], "offers": [59], "promising": [61], "solution": [62], "for": [63, 96, 162], "specialization,": [64], "existing": [65, 106], "MoE-based": [66], "methods": [67], "suffer": [68], "coarse": [70], "image-level": [71, 107], "expert": [72], "assignment": [73], "and": [74, 134], "high": [75], "communication": [76, 122, 184], "costs": [77], "parameterized": [79], "routers.": [80], "To": [81, 120], "address": [82], "these": [83], "limitations,": [84], "we": [85], "propose": [86], "TRIP,": [87], "Token-level": [89], "parameter-free": [93, 127], "routing": [94, 108, 128], "framework": [95], "FedDG,": [97], "which": [98], "treats": [99], "multiple": [100], "prompts": [101], "as": [102], "distinct": [103], "experts.": [104, 119], "Unlike": [105], "designs,": [109], "TRIP": [110, 124, 156, 178], "assigns": [111], "different": [112], "tokens": [113, 151], "within": [114], "an": [115, 158], "image": [116], "specific": [118], "ensure": [121], "efficiency,": [123], "incorporates": [125], "mechanism": [129], "based": [130], "token": [132], "clustering": [133], "optimal": [135, 180], "transport.": [136], "The": [137], "instance-specific": [138], "is": [140, 193], "then": [141], "synthesized": [142], "aggregating": [144], "experts,": [145, 164], "weighted": [146], "number": [149], "assigned": [152], "each.": [154], "Additionally,": [155], "develops": [157], "unbiased": [159], "strategy": [161], "leveraging": [165], "VLM's": [167], "zero-shot": [168], "capability.": [170], "Extensive": [171], "experiments": [172], "across": [173], "four": [174], "benchmarks": [175], "demonstrate": [176], "that": [177], "achieves": [179], "results,": [182], "only": [186], "1K": [187], "parameters": [188], "per": [189], "round.": [190], "Our": [191], "code": [192], "available": [194], "at": [195], "https://github.com/GongShuai8210/TRIP.": [196]}, "primary_topic": {"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.9932000041007996, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, "topics": [{"id": "https://openalex.org/T11307", "display_name": "Domain Adaptation and Few-Shot Learning", "score": 0.9932000041007996, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10764", "display_name": "Privacy-Preserving Technologies in Data", "score": 0.006399999838322401, "subfield": {"id": "https://openalex.org/subfields/1702", "display_name": "Artificial Intelligence"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10036", "display_name": "Advanced Neural Network Applications", "score": 9.999999747378752e-05, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}], "primary_location": {"id": "pmh:oai:arXiv.org:2504.21063", "is_oa": true, "landing_page_url": "http://arxiv.org/abs/2504.21063", "pdf_url": "https://arxiv.org/pdf/2504.21063", "source": {"id": "https://openalex.org/S4306400194", "display_name": "arXiv (Cornell University)", "issn_l": "2331-8422", "issn": ["2331-8422"], "is_oa": true, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I205783295", "host_organization_name": "Cornell University", "host_organization_lineage": ["https://openalex.org/I205783295"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "submittedVersion", "is_accepted": false, "is_published": false, "raw_source_name": null, "raw_type": "text"}, "cited_by_count": 0, "is_retracted": false}, {"id": "https://openalex.org/W4414132407", "doi": "https://doi.org/10.1101/2025.09.05.674598", "title": "MAVeRiC-AD: Mixture-of-experts Agentic Vision-Language Ensemble for Robust MRI Classification of Alzheimer\u2019s Disease", "publication_year": 2025, "publication_date": "2025-09-11", "type": "preprint", "language": "en", "authorships": [{"author_position": "first", "author": {"id": "https://openalex.org/A5089189717", "display_name": "Nikhil J. Dhinagar", "orcid": "https://orcid.org/0000-0003-2424-4854", "observed_orcids": ["https://orcid.org/0000-0003-2424-4854"]}, "institutions": [{"id": "https://openalex.org/I1174212", "display_name": "University of Southern California", "ror": "https://ror.org/03taz7m60", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1174212"]}], "countries": ["US"], "is_corresponding": true, "raw_author_name": "Nikhil J. Dhinagar", "raw_affiliation_strings": ["University of Southern California", "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Southern California", "institution_ids": ["https://openalex.org/I1174212"]}, {"raw_affiliation_string": "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA", "institution_ids": ["https://openalex.org/I1174212"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5101772582", "display_name": "Pavithra Senthilkumar", "orcid": null, "observed_orcids": []}, "institutions": [{"id": "https://openalex.org/I1174212", "display_name": "University of Southern California", "ror": "https://ror.org/03taz7m60", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1174212"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Pavithra Senthilkumar", "raw_affiliation_strings": ["University of Southern California", "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Southern California", "institution_ids": ["https://openalex.org/I1174212"]}, {"raw_affiliation_string": "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA", "institution_ids": ["https://openalex.org/I1174212"]}]}, {"author_position": "middle", "author": {"id": "https://openalex.org/A5017365268", "display_name": "Sophia I. Thomopoulos", "orcid": "https://orcid.org/0000-0002-0046-4070", "observed_orcids": ["https://orcid.org/0000-0002-0046-4070"]}, "institutions": [{"id": "https://openalex.org/I1174212", "display_name": "University of Southern California", "ror": "https://ror.org/03taz7m60", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1174212"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Sophia I. Thomopoulos", "raw_affiliation_strings": ["University of Southern California", "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Southern California", "institution_ids": ["https://openalex.org/I1174212"]}, {"raw_affiliation_string": "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA", "institution_ids": ["https://openalex.org/I1174212"]}]}, {"author_position": "last", "author": {"id": "https://openalex.org/A5100606557", "display_name": "Paul M. Thompson", "orcid": "https://orcid.org/0000-0002-4720-8867", "observed_orcids": ["https://orcid.org/0000-0002-4720-8867"]}, "institutions": [{"id": "https://openalex.org/I1174212", "display_name": "University of Southern California", "ror": "https://ror.org/03taz7m60", "country_code": "US", "type": "education", "lineage": ["https://openalex.org/I1174212"]}], "countries": ["US"], "is_corresponding": false, "raw_author_name": "Paul M. Thompson", "raw_affiliation_strings": ["University of Southern California", "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA"], "raw_orcid": null, "affiliations": [{"raw_affiliation_string": "University of Southern California", "institution_ids": ["https://openalex.org/I1174212"]}, {"raw_affiliation_string": "Imaging Genetics Center, Mark & Mary Stevens Neuroimaging & Informatics Institute, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA", "institution_ids": ["https://openalex.org/I1174212"]}]}], "abstract_inverted_index": {"Abstract": [0], "Robust": [1, 45], "classification": [2, 47], "of": [3, 48, 134, 228, 246], "Alzheimer\u2019s": [4, 49], "disease": [5], "(AD)": [6], "from": [7, 127, 169, 178], "structural": [8], "T1-weighted": [9], "MRI": [10], "(T1w)": [11], "images": [12, 177], "remains": [13], "an": [14, 51], "unmet": [15], "clinical": [16, 259], "need,": [17], "especially": [18], "when": [19], "data": [20, 240], "is": [21], "acquired": [22], "at": [23], "multiple": [24], "sites": [25], "that": [26, 54, 98, 121, 225], "differ": [27], "in": [28], "scanning": [29], "protocols": [30], "and": [31, 130, 143, 159, 182, 194, 209, 219, 244, 251], "population": [32], "demographics.": [33], "In": [34, 71], "this": [35], "paper,": [36], "we": [37, 74], "present": [38], "MAVeRiC-AD": [39, 223], "(Mixture-of-experts": [40], "Agent-guided": [41], "Vision-Language": [42], "Ensemble": [43], "for": [44, 61, 238, 256], "Imaging-based": [46], "Disease),": [50], "agentic": [52, 226], "framework": [53], "dynamically": [55], "utilizes": [56], "the": [57, 69, 101, 128, 140, 146, 154, 161, 179], "optimal": [58], "inferencing": [59], "tool": [60, 162], "radiological": [62], "queries": [63], "to": [64, 68, 153], "provide": [65], "relevant": [66], "answers": [67], "user.": [70], "our": [72], "framework,": [73], "tested": [75], "three": [76], "specialized": [77], "models": [78, 100, 212], "encoded": [79], "as": [80, 253], "callable": [81], "tools:": [82], "(1)": [83], "CNN-AD,": [84], "a": [85, 95, 110, 118, 123, 254], "3D": [86], "DenseNet": [87], "trained": [88, 200], "on": [89, 197, 202, 217, 221], "T1w": [90, 102, 176], "intensities": [91], "only;": [92], "(2)": [93], "MOE-VLM,": [94], "vision-language": [96, 211], "model": [97], "jointly": [99, 201], "with": [103, 165, 213, 234], "subject-specific": [104], "demographics": [105], "(age,": [106], "sex,": [107], "site)": [108], "via": [109], "mixture-of-experts": [111], "(MoE)": [112], "projection": [113], "head;": [114], "(3)": [115], "Retrieval": [116], "engine,": [117], "similarity-search": [119], "module": [120], "contextualizes": [122], "patient": [124], "against": [125], "others": [126], "site": [129], "reports": [131], "%": [132], "prevalence": [133], "AD.": [135], "A": [136], "light-weight": [137], "agent": [138], "analyzes": [139], "user": [141], "request": [142], "then": [144], "routes": [145], "input": [147], "(image,": [148], "or": [149], "image": [150, 248], "+": [151], "text)": [152], "appropriate": [155], "tool,": [156], "aggregates": [157], "responses": [158], "returns": [160], "response": [163], "augmented": [164], "its": [166], "confidence": [167], "derived": [168], "conformal": [170], "prediction.": [171], "Experiments": [172], "were": [173], "conducted": [174], "using": [175], "ADNI": [180, 218], "(N=4,098)": [181], "OASIS-3": [183], "(N=600)": [184], "datasets.": [185], "Single-site": [186], "training": [187], "baselines": [188], "achieved": [189], "ROC-AUC": [190, 214], "=": [191, 215], "0.79": [192], "(CNN)": [193], "0.82": [195], "(VLM)": [196], "ADNI.": [198], "When": [199], "both": [203, 207], "sites,": [204], "MOE-VLM": [205], "surpassed": [206], "image-only": [208], "standard": [210], "0.90": [216], "0.81": [220], "OASIS.": [222], "demonstrates": [224], "orchestration": [227], "complementary": [229], "expert": [230], "deep": [231], "models,": [232], "coupled": [233], "explicit": [235], "demographic": [236], "conditioning": [237], "multi-site": [239], "can": [241], "improve": [242], "robustness": [243], "interpretability": [245], "AD": [247], "analysis": [249], "pipelines": [250], "serves": [252], "blueprint": [255], "scalable,": [257], "trustworthy": [258], "AI": [260], "assistants.": [261]}, "primary_topic": {"id": "https://openalex.org/T10009", "display_name": "Dementia and Cognitive Impairment Research", "score": 0.9937000274658203, "subfield": {"id": "https://openalex.org/subfields/2738", "display_name": "Psychiatry and Mental health"}, "field": {"id": "https://openalex.org/fields/27", "display_name": "Medicine"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}, "topics": [{"id": "https://openalex.org/T10009", "display_name": "Dementia and Cognitive Impairment Research", "score": 0.9937000274658203, "subfield": {"id": "https://openalex.org/subfields/2738", "display_name": "Psychiatry and Mental health"}, "field": {"id": "https://openalex.org/fields/27", "display_name": "Medicine"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}, {"id": "https://openalex.org/T11714", "display_name": "Multimodal Machine Learning Applications", "score": 0.000699999975040555, "subfield": {"id": "https://openalex.org/subfields/1707", "display_name": "Computer Vision and Pattern Recognition"}, "field": {"id": "https://openalex.org/fields/17", "display_name": "Computer Science"}, "domain": {"id": "https://openalex.org/domains/3", "display_name": "Physical Sciences"}}, {"id": "https://openalex.org/T10086", "display_name": "Alzheimer's disease research and treatments", "score": 0.000699999975040555, "subfield": {"id": "https://openalex.org/subfields/2737", "display_name": "Physiology"}, "field": {"id": "https://openalex.org/fields/27", "display_name": "Medicine"}, "domain": {"id": "https://openalex.org/domains/4", "display_name": "Health Sciences"}}], "primary_location": {"id": "doi:10.1101/2025.09.05.674598", "is_oa": true, "landing_page_url": "https://doi.org/10.1101/2025.09.05.674598", "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2025/09/11/2025.09.05.674598.full.pdf", "source": {"id": "https://openalex.org/S4306402567", "display_name": "bioRxiv (Cold Spring Harbor Laboratory)", "issn_l": null, "issn": null, "is_oa": false, "is_in_doaj": false, "is_core": false, "listed_in": [], "host_organization": "https://openalex.org/I2750212522", "host_organization_name": "Cold Spring Harbor Laboratory", "host_organization_lineage": ["https://openalex.org/I2750212522"], "host_organization_lineage_names": [], "type": "repository"}, "license": null, "license_id": null, "version": "acceptedVersion", "is_accepted": true, "is_published": false, "raw_source_name": "bioRxiv", "raw_type": "posted-content"}, "cited_by_count": 1, "is_retracted": false}], "group_by": []}