{"id":"https://openalex.org/W7140122039","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.31","title":"KNN-SSD: Enabling Dynamic Self-Speculative Decoding via Nearest Neighbor Layer Set Optimization","display_name":"KNN-SSD: Enabling Dynamic Self-Speculative Decoding via Nearest Neighbor Layer Set Optimization","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140122039","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.31"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-eacl.31","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.31","pdf_url":"https://aclanthology.org/2026.findings-eacl.31.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: EACL 2026","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-eacl.31.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130335122","display_name":"Mingbo Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mingbo Song","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103220384","display_name":"Heming Xia","orcid":"https://orcid.org/0000-0001-5074-3441"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Heming Xia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130379886","display_name":"Jun Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jun Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5116335566","display_name":"Chak Tou Leong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chak Tou Leong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130411190","display_name":"Qiancheng Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qiancheng Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130348315","display_name":"Wenjie Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wenjie Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5058353424","display_name":"Sujian Li","orcid":"https://orcid.org/0000-0001-7493-0786"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sujian Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":18.0209,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.98803034,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"641","last_page":"655"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.7249000072479248,"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"}},"topics":[{"id":"https://openalex.org/T11181","display_name":"Advanced Data Storage Technologies","score":0.7249000072479248,"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"}},{"id":"https://openalex.org/T12326","display_name":"Network Packet Processing and Optimization","score":0.044199999421834946,"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/T11478","display_name":"Caching and Content Delivery","score":0.025599999353289604,"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/set","display_name":"Set (abstract data type)","score":0.5304999947547913},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.5171999931335449},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.48249998688697815},{"id":"https://openalex.org/keywords/k-nearest-neighbors-algorithm","display_name":"k-nearest neighbors algorithm","score":0.4645000100135803},{"id":"https://openalex.org/keywords/minification","display_name":"Minification","score":0.34439998865127563},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.33570000529289246}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5842999815940857},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5304999947547913},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5171999931335449},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.508899986743927},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.48249998688697815},{"id":"https://openalex.org/C113238511","wikidata":"https://www.wikidata.org/wiki/Q1071612","display_name":"k-nearest neighbors algorithm","level":2,"score":0.4645000100135803},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.34439998865127563},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.33570000529289246},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3073999881744385},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.30169999599456787},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.298799991607666},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2874999940395355},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.2782000005245209},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2775000035762787},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.2667999863624573}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-eacl.31","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.31","pdf_url":"https://aclanthology.org/2026.findings-eacl.31.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: EACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-eacl.31","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.31","pdf_url":"https://aclanthology.org/2026.findings-eacl.31.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: EACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3336050986","display_name":null,"funder_award_id":"62476010","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"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140122039.pdf","grobid_xml":"https://content.openalex.org/works/W7140122039.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Speculative":[0],"Decoding":[1,44],"(SD)":[2],"has":[3],"emerged":[4],"as":[5],"a":[6,30,83],"widely":[7],"used":[8],"paradigm":[9],"to":[10,49,78,82,107,130],"accelerate":[11],"the":[12,40,51,56,90],"inference":[13],"of":[14,93],"large":[15],"language":[16],"models":[17,121],"(LLMs)":[18],"without":[19],"compromising":[20],"generation":[21],"quality.It":[22],"works":[23],"by":[24],"efficiently":[25],"drafting":[26,71],"multiple":[27,123],"tokens":[28],"using":[29,39],"compact":[31],"model":[32],"and":[33,122],"then":[34],"verifying":[35],"them":[36],"in":[37,67,86,119,132],"parallel":[38],"target":[41],"LLM.Notably,":[42],"Self-Speculative":[43],"proposes":[45],"skipping":[46,74],"certain":[47],"layers":[48,111],"construct":[50],"draft":[52],"model,":[53],"which":[54],"eliminates":[55],"need":[57],"for":[58],"additional":[59],"parameters":[60],"or":[61],"training.Despite":[62],"its":[63,127],"strengths,":[64],"we":[65,96],"observe":[66],"this":[68,94],"work":[69],"that":[70,101,126],"with":[72,112],"layer":[73],"exhibits":[75],"significant":[76],"sensitivity":[77],"domain":[79,91,114],"shifts,":[80],"leading":[81],"substantial":[84],"drop":[85],"acceleration":[87],"performance.To":[88],"enhance":[89],"generalizability":[92],"paradigm,":[95],"introduce":[97],"KNN-SSD,":[98],"an":[99],"algorithm":[100,118],"leverages":[102],"K-Nearest":[103],"Neighbor":[104],"(KNN)":[105],"search":[106],"match":[108],"different":[109],"skipped":[110],"various":[113,120],"inputs.We":[115],"evaluated":[116],"our":[117],"tasks,":[124],"observing":[125],"application":[128],"leads":[129],"1.31.6speedup":[131],"LLM":[133],"inference.":[134]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-03-24T00:00:00"}