{"id":"https://openalex.org/W7153184966","doi":"https://doi.org/10.1109/access.2026.3680126","title":"QubitCache: Quantum-Inspired Probabilistic Attention Preservation for KV-Cache Compression","display_name":"QubitCache: Quantum-Inspired Probabilistic Attention Preservation for KV-Cache Compression","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7153184966","doi":"https://doi.org/10.1109/access.2026.3680126"},"language":"en","primary_location":{"id":"doi:10.1109/access.2026.3680126","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3680126","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1109/access.2026.3680126","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5058572236","display_name":"Jieui Kang","orcid":"https://orcid.org/0009-0000-7691-0930"},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jieui Kang","raw_affiliation_strings":["Artificial Intelligence Convergence, Ewha Womans University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0009-0000-7691-0930","affiliations":[{"raw_affiliation_string":"Artificial Intelligence Convergence, Ewha Womans University, Seoul, South Korea","institution_ids":["https://openalex.org/I138925566"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121959511","display_name":"Jaeyoung Choi","orcid":null},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jaeyoung Choi","raw_affiliation_strings":["Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea","institution_ids":["https://openalex.org/I138925566"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133386587","display_name":"Wonhui Noh","orcid":null},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Wonhui Noh","raw_affiliation_strings":["Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0009-0002-3936-2896","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea","institution_ids":["https://openalex.org/I138925566"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5013362846","display_name":"Jaehyeong Sim","orcid":"https://orcid.org/0000-0001-8722-8486"},"institutions":[{"id":"https://openalex.org/I138925566","display_name":"Ewha Womans University","ror":"https://ror.org/053fp5c05","country_code":"KR","type":"education","lineage":["https://openalex.org/I138925566"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jaehyeong Sim","raw_affiliation_strings":["Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea"],"raw_orcid":"https://orcid.org/0000-0001-8722-8486","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Engineering, Ewha Womans University, Seoul, South Korea","institution_ids":["https://openalex.org/I138925566"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I138925566"],"apc_list":{"value":1850,"currency":"USD","value_usd":1850},"apc_paid":{"value":1850,"currency":"USD","value_usd":1850},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.41818481,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"14","issue":null,"first_page":"56679","last_page":"56693"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10682","display_name":"Quantum Computing Algorithms and Architecture","score":0.5475999712944031,"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/T10682","display_name":"Quantum Computing Algorithms and Architecture","score":0.5475999712944031,"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/T14347","display_name":"Big Data and Digital Economy","score":0.0568000003695488,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.03889999911189079,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/probabilistic-logic","display_name":"Probabilistic logic","score":0.6754999756813049},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.5990999937057495},{"id":"https://openalex.org/keywords/data-compression","display_name":"Data compression","score":0.5679000020027161},{"id":"https://openalex.org/keywords/compression-artifact","display_name":"Compression artifact","score":0.2549000084400177},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.24560000002384186}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7739999890327454},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.6754999756813049},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.5990999937057495},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.5679000020027161},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4672999978065491},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3240000009536743},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2644999921321869},{"id":"https://openalex.org/C57654395","wikidata":"https://www.wikidata.org/wiki/Q1097775","display_name":"Compression artifact","level":5,"score":0.2549000084400177},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.24560000002384186},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.22840000689029694}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/access.2026.3680126","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3680126","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:fb66f9ac3cbd42e68305e89de37de011","is_oa":true,"landing_page_url":"https://doaj.org/article/fb66f9ac3cbd42e68305e89de37de011","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"IEEE Access, Vol 14, Pp 56679-56693 (2026)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1109/access.2026.3680126","is_oa":true,"landing_page_url":"https://doi.org/10.1109/access.2026.3680126","pdf_url":null,"source":{"id":"https://openalex.org/S2485537415","display_name":"IEEE Access","issn_l":"2169-3536","issn":["2169-3536"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Access","raw_type":"journal-article"},"sustainable_development_goals":[{"score":0.6596503257751465,"id":"https://metadata.un.org/sdg/13","display_name":"Climate action"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"},{"id":"https://openalex.org/F4320335489","display_name":"Institute for Information and Communications Technology Promotion","ror":"https://ror.org/01g0hqq23"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"model":[2],"inference":[3],"faces":[4],"a":[5,44,74,94,120],"critical":[6,97],"memory":[7,21,29,255],"bottleneck":[8],"from":[9,70],"KV":[10,48],"cache,":[11],"which":[12,148,270],"grows":[13],"linearly":[14],"with":[15,263],"sequence":[16],"length":[17],"and":[18,174,187,235,246,296,304],"dominates":[19],"GPU":[20],"during":[22,178],"long-context":[23,229],"generation.":[24],"Existing":[25],"compression":[26,50,54],"methods":[27,248],"reduce":[28],"through":[30,218],"token":[31,57],"eviction":[32],"but":[33],"irreversibly":[34],"discard":[35],"attention":[36,52,85,108,142,155,210],"relationships":[37],"essential":[38],"for":[39,197],"complex":[40],"reasoning.We":[41],"present":[42],"QubitCache,":[43],"framework":[45],"that":[46,131,166,191,280],"reframes":[47],"cache":[49],"as":[51,73],"distribution":[53,77,169],"rather":[55,220],"than":[56,221],"selection.":[58],"Our":[59],"approach":[60,302],"employs":[61],"aquantum-inspiredmathematical":[62],"formalism":[63],"(using":[64],"the":[65,132,139,167,182,192,198,236],"hierarchical":[66,116],"amplitude":[67,117],"encoding":[68],"structure":[69],"quantum":[71,81],"computing":[72],"classical":[75,105,242],"probability":[76],"compressor,":[78],"without":[79],"requiring":[80],"hardware)":[82],"to":[83,145,153,207],"represent":[84],"weight":[86],"distributions":[87,109,204],"in":[88,103],"logarithmic":[89],"parameter":[90,122],"space.":[91],"QubitCache":[92,226,251],"introduces":[93],"hybrid":[95],"architecture:":[96],"tokens":[98,112,217],"(15%":[99],"of":[100,141,162,184,215,260,266,287],"sequence)":[101],"remain":[102],"full-precision":[104],"storage,":[106],"while":[107,257],"over":[110],"remaining":[111],"are":[113,205],"compressed":[114,146,168,203,216],"using":[115],"encoding,":[118],"achieving":[119],"57\u00d7":[121],"reduction":[123,256],"per":[124],"segment.":[125],"We":[126,224],"provide":[127,188,208],"formal":[128],"analysis":[129,278,291],"showing":[130],"pipeline":[133],"output":[134],"error":[135],"is":[136,149,165,170,175,195,271],"bounded":[137],"by":[138],"fraction":[140],"mass":[143],"assigned":[144],"tokens,":[147],"small":[150],"(\u226415%)":[151],"due":[152],"inherent":[154],"sparsity":[156],"(Theorem":[157],"1).":[158],"An":[159],"important":[160],"property":[161],"our":[163],"design":[164],"fixed":[171],"after":[172],"prefill":[173],"therefore":[176],"query-agnostic":[177],"decoding;":[179],"we":[180],"analyze":[181],"implications":[183],"this":[185,301],"trade-off":[186],"empirical":[189],"evidence":[190],"resulting":[193],"\u03bbmassstability":[194],"sufficient":[196],"tasks":[199,234],"evaluated.":[200],"During":[201],"inference,":[202],"reconstructed":[206],"soft":[209],"weights,":[211],"maintaining":[212],"contextual":[213],"influence":[214],"probabilistic":[219],"binary":[222],"decisions.":[223],"evaluate":[225],"on":[227,268],"four":[228],"models":[230],"across":[231],"sixteen":[232],"LongBench":[233],"RULER":[237],"benchmark,":[238],"comparing":[239],"against":[240],"both":[241],"baselines":[243],"(H2O,":[244],"SnapKV)":[245],"recent":[247],"(MiniKV,":[249],"Compactor).":[250],"achieves":[252],"approximately":[253],"6.7\u00d7":[254],"retaining":[258],"91\u201399%":[259],"full-KV":[261],"performance,":[262],"an":[264],"average":[265],"74.8":[267],"RULER,":[269],"10.5":[272],"points":[273],"above":[274],"token-eviction":[275],"baselines.":[276],"Latency":[277],"confirms":[279],"per-token":[281],"decode":[282],"overhead":[283],"remains":[284],"within":[285],"2\u00d7":[286],"eviction-based":[288],"methods.":[289],"Comprehensive":[290],"including":[292],"sensitivity,":[293],"quantization":[294],"integration,":[295],"failure":[297],"cases":[298],"characterizes":[299],"when":[300],"succeeds":[303],"where":[305],"it":[306],"has":[307],"limitations.":[308]},"counts_by_year":[],"updated_date":"2026-07-29T09:40:50.615796","created_date":"2026-04-11T00:00:00"}