{"id":"https://openalex.org/W4416037098","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.63","title":"SAKI-RAG: Mitigating Context Fragmentation in Long-Document RAG via Sentence-level Attention Knowledge Integration","display_name":"SAKI-RAG: Mitigating Context Fragmentation in Long-Document RAG via Sentence-level Attention Knowledge Integration","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416037098","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.63"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.emnlp-main.63","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.63","pdf_url":"https://aclanthology.org/2025.emnlp-main.63.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":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.emnlp-main.63.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5112650930","display_name":"W. Tao","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Wenyu Tao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101395450","display_name":"Xiaofen Xing","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaofen Xing","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087980711","display_name":"Zeliang Li","orcid":"https://orcid.org/0000-0003-2757-3688"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zeliang Li","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5033271721","display_name":"Xiangmin Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiangmin Xu","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5112650930"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.18461017,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1195","last_page":"1213"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.633899986743927,"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":0.633899986743927,"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":0.08649999648332596,"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":0.07980000227689743,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5723999738693237},{"id":"https://openalex.org/keywords/fragmentation","display_name":"Fragmentation (computing)","score":0.5550000071525574},{"id":"https://openalex.org/keywords/knowledge-based-systems","display_name":"Knowledge-based systems","score":0.2574999928474426}],"concepts":[{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5723999738693237},{"id":"https://openalex.org/C191015642","wikidata":"https://www.wikidata.org/wiki/Q1132459","display_name":"Fragmentation (computing)","level":2,"score":0.5550000071525574},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5357000231742859},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.36640000343322754},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.3075999915599823},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.27250000834465027},{"id":"https://openalex.org/C115925183","wikidata":"https://www.wikidata.org/wiki/Q1412694","display_name":"Knowledge-based systems","level":2,"score":0.2574999928474426},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2547000050544739},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.25369998812675476},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.2433999925851822}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.emnlp-main.63","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.63","pdf_url":"https://aclanthology.org/2025.emnlp-main.63.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":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.emnlp-main.63","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.63","pdf_url":"https://aclanthology.org/2025.emnlp-main.63.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":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416037098.pdf","grobid_xml":"https://content.openalex.org/works/W4416037098.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Traditional":[0],"Retrieval-Augmented":[1,71],"Generation":[2],"(RAG)":[3],"frameworks":[4,21,137],"often":[5],"segment":[6],"documents":[7],"into":[8],"larger":[9],"chunks":[10,57,106],"to":[11,27,37,100,134],"preserve":[12],"contextual":[13,115],"coherence,":[14],"inadvertently":[15],"introducing":[16],"redundant":[17],"noise.Recent":[18],"advanced":[19],"RAG":[20,136],"have":[22],"shifted":[23],"toward":[24],"finer-grained":[25],"chunking":[26,34],"improve":[28],"precision.However,":[29],"in":[30,138],"long-document":[31,139],"scenarios,":[32,141],"such":[33],"methods":[35],"lead":[36],"fragmented":[38],"contexts,":[39],"isolated":[40],"chunk":[41],"semantics,":[42],"and":[43,92,102,114,123,131],"broken":[44],"inter-chunk":[45],"relationships,":[46,91],"making":[47],"cross-paragraph":[48],"retrieval":[49,140],"particularly":[50],"challenging.To":[51],"address":[52],"this":[53],"challenge,":[54],"maintaining":[55],"granular":[56],"while":[58,142],"recovering":[59],"their":[60],"intrinsic":[61],"semantic":[62,112],"connections,":[63],"we":[64],"propose":[65],"SAKI-RAG":[66,127],"(Sentence-level":[67],"Attention":[68],"Knowledge":[69],"Integration":[70],"Generation).Our":[72],"framework":[73],"introduces":[74],"two":[75],"core":[76],"components:(1)":[77],"the":[78,94,104,108],"SentenceAttnLinker,":[79],"which":[80,97],"constructs":[81],"a":[82],"semantically":[83],"enriched":[84],"knowledge":[85],"repository":[86],"by":[87],"modeling":[88],"inter-sentence":[89],"attention":[90],"(2)":[93],"Dual-Axis":[95],"Retriever,":[96],"is":[98],"designed":[99],"expand":[101],"filter":[103],"candidate":[105],"from":[107],"dual":[109],"dimensions":[110],"of":[111],"similarity":[113],"relevance.Experimental":[116],"results":[117],"across":[118],"four":[119],"datasets-Dragonball,":[120],"SQUAD,":[121],"NFCORPUS,":[122],"SCI-DOCS":[124],"demonstrate":[125],"that":[126],"achieves":[128],"better":[129],"recall":[130],"precision":[132],"compared":[133],"other":[135],"also":[143],"exhibiting":[144],"higher":[145],"information":[146],"efficiency.":[147]},"counts_by_year":[],"updated_date":"2026-03-13T14:20:09.374765","created_date":"2025-11-08T00:00:00"}