{"id":"https://openalex.org/W7224599385","doi":"https://doi.org/10.18653/2026.emnlp-main.229","title":"EarlySelectKV: Anticipatory KV Selection via Cross-Layer Query Proxies","display_name":"EarlySelectKV: Anticipatory KV Selection via Cross-Layer Query Proxies","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7224599385","doi":"https://doi.org/10.18653/2026.emnlp-main.229"},"language":"en","primary_location":{"id":"doi:10.18653/2026.emnlp-main.229","doi":"https://doi.org/10.18653/2026.emnlp-main.229","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.229","pdf_url":"https://aclanthology.org/2026.emnlp-main.229.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","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":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2026.emnlp-main.229.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5160092295","display_name":"Yuhao Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"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":"Yuhao Liu","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","University of Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160146784","display_name":"Yuehao Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"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":"Yuehao Li","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","University of Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160147019","display_name":"Linghe Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"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":"Linghe Zhang","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","University of Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160114180","display_name":"Gaobin Huang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"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":"Gaobin Huang","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","University of Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160130257","display_name":"Junran Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"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":"Junran Li","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","University of Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]},{"raw_affiliation_string":"University of Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160141437","display_name":"Hongchang Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hongchang Wang","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5160068946","display_name":"Haifang Jian","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210149211","display_name":"Institute of Semiconductors","ror":"https://ror.org/048dd0611","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haifang Jian","raw_affiliation_strings":["Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Laboratory of Solid State Optoelectronics Information Technology , Institute of Semiconductors , Chinese Academy of Sciences , Beijing , China","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210149211"]}]}],"institutions":[],"countries_distinct_count":1,"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":0.77118064,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"5272","last_page":"5291"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.5586000084877014,"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":0.5586000084877014,"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":0.2808000147342682,"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":0.07010000199079514,"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/query-routing","display_name":"query routing","score":0.640999972820282},{"id":"https://openalex.org/keywords/long-context-decoding","display_name":"long-context decoding","score":0.781000018119812},{"id":"https://openalex.org/keywords/inference-throughput","display_name":"inference throughput","score":0.5450000166893005},{"id":"https://openalex.org/keywords/gpu-memory-reduction","display_name":"GPU memory reduction","score":0.9520000219345093}],"concepts":[{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5874999761581421},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5742999911308289},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39890000224113464},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.35109999775886536},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.2930000126361847},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.2808000147342682},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.25529998540878296},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.24809999763965607},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.24300000071525574},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.23819999396800995}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/2026.emnlp-main.229","doi":"https://doi.org/10.18653/2026.emnlp-main.229","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.229","pdf_url":"https://aclanthology.org/2026.emnlp-main.229.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","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":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2611.01715","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.01715","pdf_url":"https://arxiv.org/pdf/2611.01715","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"},{"id":"doi:10.48550/arxiv.2611.01715","doi":"https://doi.org/10.48550/arxiv.2611.01715","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.01715","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,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/2026.emnlp-main.229","doi":"https://doi.org/10.18653/2026.emnlp-main.229","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.229","pdf_url":"https://aclanthology.org/2026.emnlp-main.229.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","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":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G2729455034","display_name":null,"funder_award_id":"2024YFE0210600","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program 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/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7224599385.pdf","grobid_xml":"https://content.openalex.org/works/W7224599385.grobid-xml"},"referenced_works_count":6,"referenced_works":["https://openalex.org/W2988394319","https://openalex.org/W4387561528","https://openalex.org/W4396815229","https://openalex.org/W4401306886","https://openalex.org/W4405655184","https://openalex.org/W7147287851"],"related_works":["https://openalex.org/W7224433058","https://openalex.org/W7224247817","https://openalex.org/W7166882967","https://openalex.org/W7224245099","https://openalex.org/W7166821537","https://openalex.org/W7224168732","https://openalex.org/W7211956120","https://openalex.org/W7220792417","https://openalex.org/W7224271966","https://openalex.org/W7224446656"],"abstract_inverted_index":{"Query-conditioned":[0],"sparse":[1,146],"KV":[2,15,57,142],"selection":[3,144],"reduces":[4,124],"long-context":[5,147],"inference":[6],"cost":[7],"by":[8,90,119,127],"attending":[9],"to":[10,54,121],"a":[11,43,108],"subset":[12,143],"of":[13,77,129],"historical":[14],"entries":[16],"at":[17],"each":[18],"decoding":[19],"step.Existing":[20],"selectors":[21],"must":[22],"wait":[23],"for":[24,65],"the":[25,37,61,81,84,130],"target-layer":[26,63],"query":[27,64],"before":[28],"scoring":[29],"candidates":[30],"and":[31,71,100,123],"generating":[32],"indices,":[33],"placing":[34],"routing":[35,135],"on":[36,93,98,104],"immediate":[38],"pre-attention":[39],"path.We":[40],"introduce":[41],"EARLYSELECTKV,":[42],"trainingfree":[44],"transformation":[45],"that":[46],"uses":[47],"proxy":[48],"queries":[49],"from":[50],"earlier":[51],"hidden":[52],"states":[53],"precompute":[55],"selected":[56],"indices":[58],"while":[59],"retaining":[60],"real":[62],"final":[66],"attention.Tests":[67],"with":[68,111],"RocketKV,":[69],"Quest,":[70],"Loki":[72],"show":[73],"no":[74],"consistent":[75],"pattern":[76],"quality":[78],"degradation":[79],"across":[80],"evaluated":[82],"settings.In":[83,107],"primary":[85],"Rock-etKV":[86],"evaluation,":[87],"ES-RocketKV":[88,114],"differs":[89],"+0.018":[91],"points":[92,97,103],"LongBench,":[94],"approximately":[95],"-0.01":[96],"RULER,":[99],"+0.25":[101],"percentage":[102],"representative":[105],"NIAH/Passkey":[106],"TensorRT-LLM":[109],"deployment":[110],"CPU-offloaded":[112],"routing,":[113],"improves":[115],"end-to-end":[116],"output":[117],"throughput":[118],"up":[120],"12.4%":[122],"GPU-resident":[125],"memory":[126],"11.1%":[128],"combined":[131],"KV/K-summary":[132],"pool.This":[133],"makes":[134],"timing":[136],"an":[137],"explicit":[138],"optimization":[139],"target":[140],"alongside":[141],"in":[145],"decoding.":[148]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-10-30T00:00:00"}