{"id":"https://openalex.org/W7166850384","doi":"https://doi.org/10.18653/v1/2026.findings-acl.201","title":"HqeKV: Towards Hybrid Quantization and Eviction for KV Cache in Long-Context LLM Inference","display_name":"HqeKV: Towards Hybrid Quantization and Eviction for KV Cache in Long-Context LLM Inference","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166850384","doi":"https://doi.org/10.18653/v1/2026.findings-acl.201"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.201","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.201","pdf_url":"https://aclanthology.org/2026.findings-acl.201.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":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.201.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139747691","display_name":"He Wang","orcid":null},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"He Wang","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139842080","display_name":"Yu Gu","orcid":null},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yu Gu","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139787399","display_name":"Fangfang Li","orcid":null},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Fangfang Li","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139737484","display_name":"Zhigang Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210106808","display_name":"Guangzhou Institute of Advanced Technology","ror":"https://ror.org/01vcw4681","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210106808"]},{"id":"https://openalex.org/I59028903","display_name":"Ocean University of China","ror":"https://ror.org/04rdtx186","country_code":"CN","type":"education","lineage":["https://openalex.org/I59028903"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhigang Wang","raw_affiliation_strings":["Cyberspace Institute of Advanced Technology , Guangzhou University ,","Faculty of Information Science and Engineering , Ocean University of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Cyberspace Institute of Advanced Technology , Guangzhou University ,","institution_ids":["https://openalex.org/I4210106808"]},{"raw_affiliation_string":"Faculty of Information Science and Engineering , Ocean University of China","institution_ids":["https://openalex.org/I59028903"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100769917","display_name":"Zhenyu Liu","orcid":"https://orcid.org/0000-0002-4205-6067"},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhenghao Liu","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139811044","display_name":"Ning Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210106808","display_name":"Guangzhou Institute of Advanced Technology","ror":"https://ror.org/01vcw4681","country_code":"CN","type":"education","lineage":["https://openalex.org/I4210106808"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ning Wang","raw_affiliation_strings":["Cyberspace Institute of Advanced Technology , Guangzhou University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Cyberspace Institute of Advanced Technology , Guangzhou University ,","institution_ids":["https://openalex.org/I4210106808"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139825887","display_name":"Xiaohua Li","orcid":null},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xiaohua Li","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5139767265","display_name":"Ge Yu","orcid":null},"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"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ge Yu","raw_affiliation_strings":["School of Computer Science and Engineering , Northeastern University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering , Northeastern University ,","institution_ids":["https://openalex.org/I12912129"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"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":0.79419643,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"4138","last_page":"4153"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.08879999816417694,"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":0.08879999816417694,"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/T12326","display_name":"Network Packet Processing and Optimization","score":0.08789999783039093,"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":0.0877000018954277,"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/quantization","display_name":"Quantization (signal processing)","score":0.5295000076293945},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5009999871253967},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.48179998993873596},{"id":"https://openalex.org/keywords/cpu-cache","display_name":"CPU cache","score":0.25699999928474426},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.2401999980211258}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6370999813079834},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.5295000076293945},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5009999871253967},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.48179998993873596},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.46480000019073486},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.29499998688697815},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28760001063346863},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.25699999928474426},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2401999980211258},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.23389999568462372}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.201","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.201","pdf_url":"https://aclanthology.org/2026.findings-acl.201.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"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.201","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.201","pdf_url":"https://aclanthology.org/2026.findings-acl.201.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/G3791426778","display_name":null,"funder_award_id":"62572108","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"},{"id":"https://openalex.org/F4320337111","display_name":"Basic and Applied Basic Research Foundation of Guangdong Province","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166850384.pdf","grobid_xml":"https://content.openalex.org/works/W7166850384.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"autoregressive":[1],"inference":[2,61],"in":[3,26],"large":[4],"language":[5],"models":[6],"requires":[7],"repeated":[8],"computation":[9],"across":[10,142],"transformer":[11],"layers.While":[12],"caching":[13],"intermediate":[14],"key-value":[15],"(KV)":[16],"pairs":[17],"eliminates":[18],"redundancy,":[19],"it":[20],"introduces":[21],"severe":[22],"memory":[23,162],"overhead,":[24],"particularly":[25],"long-context":[27],"settings.Most":[28],"existing":[29],"cache":[30],"compression":[31,52,70,80,99],"methods":[32],"operate":[33],"solely":[34],"on":[35,41,73],"either":[36],"quantization":[37,75,144],"or":[38],"eviction,":[39,77],"based":[40],"importance":[42,56,88,120,126],"estimation":[43],"of":[44,89],"cached":[45,90,103],"data.However,":[46],"they":[47],"are":[48],"limited":[49],"by":[50],"coarse":[51],"choices":[53],"and":[54,76],"inaccurate":[55],"assessment,":[57],"leading":[58],"to":[59,85,122,147],"suboptimal":[60],"quality.To":[62],"address":[63],"this,":[64],"we":[65],"propose":[66],"HqeKV,":[67],"a":[68,117],"hybrid":[69],"framework":[71],"built":[72],"both":[74],"offering":[78],"finer-grained":[79],"options":[81],"that":[82,130,154],"adapt":[83],"smoothly":[84],"the":[86,97,131,160],"varying":[87],"KV":[91],"pairs.An":[92],"integrated":[93],"optimizer":[94,132],"automatically":[95],"selects":[96],"best":[98],"action":[100],"for":[101],"each":[102],"element,":[104],"maximizing":[105],"quality":[106,150,158],"while":[107],"insulating":[108],"end-users":[109],"from":[110],"tedious":[111],"low-level":[112],"tuning":[113],"details.We":[114],"further":[115,148],"design":[116],"joint":[118],"K-V":[119],"metric":[121],"provide":[123],"more":[124],"accurate":[125],"assessment":[127],"results":[128],"so":[129],"can":[133],"make":[134],"smarter":[135],"decisions.Additionally,":[136],"HqeKV":[137,155],"supports":[138],"flexible":[139],"conversion":[140],"policies":[141],"multiple":[143],"precision":[145],"levels,":[146],"reduce":[149],"degradation.Extensive":[151],"experiments":[152],"show":[153],"improves":[156],"output":[157],"under":[159],"same":[161],"constraints,":[163],"outperforming":[164],"stateof-the-art":[165],"alternatives.":[166]},"counts_by_year":[],"updated_date":"2026-08-18T07:49:30.821534","created_date":"2026-07-02T00:00:00"}
