{"id":"https://openalex.org/W4416432182","doi":"https://doi.org/10.18653/v1/2026.findings-acl.695","title":"Optimizing Native Sparse Attention with Latent Attention and Local Global Alternating Strategies","display_name":"Optimizing Native Sparse Attention with Latent Attention and Local Global Alternating Strategies","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W4416432182","doi":"https://doi.org/10.18653/v1/2026.findings-acl.695"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.695","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.695","pdf_url":"https://aclanthology.org/2026.findings-acl.695.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.695.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104232073","display_name":"Yuxuan Hu","orcid":"https://orcid.org/0009-0004-3558-4584"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuxuan Hu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5052925924","display_name":"J TAN","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jianchao Tan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100349803","display_name":"Jiaqi Zhang","orcid":"https://orcid.org/0000-0002-3639-2603"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiaqi Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5120462631","display_name":"Wen Zan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wen Zan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104260560","display_name":"Pingwei Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pingwei Sun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038813094","display_name":"Yifan Lu","orcid":"https://orcid.org/0000-0002-0661-3372"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yifan Lu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5143226526","display_name":"Xunliang Cai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xunliang Cai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5143337425","display_name":"Jing Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jing Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.02662019,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14209","last_page":"14220"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.24310000240802765,"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":0.24310000240802765,"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":0.08290000259876251,"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":0.06849999725818634,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.5304999947547913},{"id":"https://openalex.org/keywords/visual-attention","display_name":"Visual attention","score":0.4327999949455261},{"id":"https://openalex.org/keywords/attention-network","display_name":"Attention network","score":0.3424000144004822},{"id":"https://openalex.org/keywords/selective-attention","display_name":"Selective attention","score":0.3370000123977661},{"id":"https://openalex.org/keywords/latent-variable","display_name":"Latent variable","score":0.32190001010894775}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6830999851226807},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5304999947547913},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.507099986076355},{"id":"https://openalex.org/C2986089797","wikidata":"https://www.wikidata.org/wiki/Q6501338","display_name":"Visual attention","level":3,"score":0.4327999949455261},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4072999954223633},{"id":"https://openalex.org/C2993807640","wikidata":"https://www.wikidata.org/wiki/Q103709453","display_name":"Attention network","level":2,"score":0.3424000144004822},{"id":"https://openalex.org/C2989146674","wikidata":"https://www.wikidata.org/wiki/Q4818236","display_name":"Selective attention","level":3,"score":0.3370000123977661},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.32679998874664307},{"id":"https://openalex.org/C51167844","wikidata":"https://www.wikidata.org/wiki/Q4422623","display_name":"Latent variable","level":2,"score":0.32190001010894775},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.2623000144958496}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.695","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.695","pdf_url":"https://aclanthology.org/2026.findings-acl.695.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2511.00819","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2511.00819","pdf_url":"https://arxiv.org/pdf/2511.00819","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,"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"},{"id":"doi:10.48550/arxiv.2511.00819","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2511.00819","pdf_url":null,"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,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.695","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.695","pdf_url":"https://aclanthology.org/2026.findings-acl.695.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1910011397","display_name":null,"funder_award_id":"U23A20299","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G2079473452","display_name":null,"funder_award_id":"U24B20144","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G3972953005","display_name":null,"funder_award_id":"62322214","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416432182.pdf","grobid_xml":"https://content.openalex.org/works/W4416432182.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"In":[0],"this":[1],"work,":[2],"we":[3,54],"conduct":[4],"a":[5],"systematic":[6],"analysis":[7],"of":[8,44],"Native":[9],"Sparse":[10],"Attention":[11,61,71,81],"(NSA)":[12],"and":[13,29,47,75,97,111,122,130],"propose":[14],"targeted":[15],"improvements":[16],"that":[17,24,62],"enhance":[18],"longcontext":[19],"modeling.A":[20],"key":[21],"insight":[22],"is":[23,66],"alternating":[25],"between":[26],"local":[27],"(sliding-window)":[28],"global":[30],"(compression/selective)":[31],"attention":[32,121,125],"across":[33],"layers,":[34],"rather":[35],"than":[36],"using":[37],"fixed":[38],"patterns,":[39],"enables":[40],"more":[41],"effective":[42],"propagation":[43],"long-range":[45],"dependencies":[46],"substantially":[48],"boosts":[49],"performance":[50],"on":[51,101,109],"long-sequence":[52],"tasks.Meanwhile,":[53],"further":[55],"refine":[56],"NSA's":[57],"branches":[58,77],"with":[59,68],"Latent":[60,70,80],"the":[63,93],"sliding-window":[64],"branch":[65],"enhanced":[67],"Multi-head":[69],"(MLA)":[72],"while":[73,91],"compression":[74],"selective":[76],"adopt":[78],"Grouphead":[79],"(GLA).These":[82],"changes":[83],"reduce":[84],"KV-cache":[85],"memory":[86],"by":[87],"50%":[88],"versus":[89],"NSA":[90],"improving":[92],"model's":[94],"common-sense":[95,128],"reasoning":[96,129],"long-text":[98],"understanding":[99,132],"capabilities.Experiments":[100],"models":[102],"from":[103],"340M":[104],"to":[105],"1.3B":[106],"parameters":[107],"(trained":[108],"15B":[110],"100B":[112],"tokens)":[113],"show":[114],"our":[115],"method":[116],"matches":[117],"or":[118],"exceeds":[119],"full":[120],"native":[123],"sparse":[124],"in":[126],"both":[127],"long-context":[131],"tasks.":[133]},"counts_by_year":[],"updated_date":"2026-08-05T07:39:15.569665","created_date":"2025-11-06T00:00:00"}