{"id":"https://openalex.org/W4416034604","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.515","title":"FIER: Fine-Grained and Efficient KV Cache Retrieval for Long-context LLM Inference","display_name":"FIER: Fine-Grained and Efficient KV Cache Retrieval for Long-context LLM Inference","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034604","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.515"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.515","pdf_url":"https://aclanthology.org/2025.findings-emnlp.515.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: EMNLP 2025","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.515.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100735740","display_name":"Dongwei Wang","orcid":"https://orcid.org/0000-0002-0943-0088"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dongwei Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101923344","display_name":"Zijie Liu","orcid":"https://orcid.org/0009-0002-4288-4388"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zijie Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102991129","display_name":"Song Wang","orcid":"https://orcid.org/0000-0001-7986-9806"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102904368","display_name":"Yu\u2010Xin Ren","orcid":"https://orcid.org/0000-0002-3047-4923"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuxin Ren","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018946131","display_name":"Jianing Deng","orcid":"https://orcid.org/0000-0001-7307-6629"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jianing Deng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066534595","display_name":"Jingtong Hu","orcid":"https://orcid.org/0000-0003-4029-4034"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jingtong Hu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017041036","display_name":"Tianlong Chen","orcid":"https://orcid.org/0009-0009-1397-9266"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tianlong Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5076154259","display_name":"Huanrui Yang","orcid":"https://orcid.org/0000-0002-3384-4512"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huanrui Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"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":0.15507531,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"9702","last_page":"9713"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.24269999563694,"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":0.24269999563694,"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/T12326","display_name":"Network Packet Processing and Optimization","score":0.22660000622272491,"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.09650000184774399,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4578999876976013},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.39070001244544983},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.23909999430179596},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.2313999980688095}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7409999966621399},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4578999876976013},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40119999647140503},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.39070001244544983},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.34380000829696655},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.31310001015663147},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.24879999458789825},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.23909999430179596},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2313999980688095},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.21930000185966492}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.515","pdf_url":"https://aclanthology.org/2025.findings-emnlp.515.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: EMNLP 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.515","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.515","pdf_url":"https://aclanthology.org/2025.findings-emnlp.515.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: EMNLP 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3578496566","display_name":null,"funder_award_id":"CNS-2122320","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G5527412688","display_name":"Collaborative Research: CNS Core:Small:IMPERIAL: In-Memory Processing Enhanced Racetrack Inspired by Accessing Laterally","funder_award_id":"2133267","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6077280891","display_name":"Collaborative Research: FuSe: R3AP: Retunable, Reconfigurable, Racetrack-Memory Acceleration Platform","funder_award_id":"2328972","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6319575343","display_name":"Collaborative Research: CNS Core: Small: Towards Unsupervised Learning on Resource Constrained Edge Devices with Novel Statistical Contrastive Learning Scheme","funder_award_id":"2122320","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320307791","display_name":"Cisco Systems","ror":"https://ror.org/03yt1ez60"},{"id":"https://openalex.org/F4320310160","display_name":"University of Arizona","ror":"https://ror.org/03m2x1q45"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416034604.pdf","grobid_xml":"https://content.openalex.org/works/W4416034604.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"Key-Value":[1],"(KV)":[2],"cache":[3,29,53,96,128],"reading":[4],"latency":[5,136],"increases":[6],"significantly":[7],"with":[8],"context":[9],"lengths,":[10],"hindering":[11],"the":[12,65,105,127],"efficiency":[13],"of":[14,27,107,126],"long-context":[15,132],"LLM":[16],"inference.To":[17],"address":[18],"this,":[19],"previous":[20],"works":[21],"propose":[22,89],"retaining":[23],"a":[24,91],"small":[25],"fraction":[26],"KV":[28,35,44,72,95,121],"based":[30],"on":[31],"token":[32],"importance.For":[33],"example,":[34],"eviction":[36],"uses":[37,99],"static":[38],"heuristics":[39],"to":[40,103,139],"retain":[41],"tokens,":[42],"while":[43],"retrieval":[45,73],"dynamically":[46],"selects":[47],"queryrelevant":[48],"tokens":[49,59,81],"for":[50],"more":[51],"adaptive":[52],"management.However,":[54],"we":[55,88],"observe":[56],"that":[57,117],"important":[58],"are":[60],"often":[61],"sparsely":[62],"distributed":[63],"across":[64,130],"long":[66],"context.This":[67],"sparsity":[68],"makes":[69],"existing":[70],"page-level":[71],"inaccurate,":[74],"as":[75],"each":[76,108],"page":[77],"may":[78],"include":[79],"irrelevant":[80],"and":[82,93,113],"miss":[83],"critical":[84],"ones.In":[85],"this":[86],"work,":[87],"Fier,":[90],"Fine-Grained":[92],"Efficient":[94],"Retrieval":[97],"method.Fier":[98],"1-bit":[100],"quantized":[101],"keys":[102],"estimate":[104],"importance":[106],"token,":[109],"resulting":[110],"in":[111],"efficient":[112],"precise":[114],"retrieval.Experiments":[115],"show":[116],"Fier":[118],"matches":[119],"full":[120],"performance":[122],"using":[123],"only":[124],"11%":[125],"budget":[129],"various":[131],"tasks,":[133],"reducing":[134],"decoding":[135],"by":[137],"1.2":[138],"1.5.":[140]},"counts_by_year":[],"updated_date":"2026-06-22T08:00:12.763002","created_date":"2025-11-08T00:00:00"}