{"id":"https://openalex.org/W7137858023","doi":"https://doi.org/10.1609/aaai.v40i31.39789","title":"DIAA: A Decoding-Efficient Inference Acceleration Approach for On-Device Large Language Models","display_name":"DIAA: A Decoding-Efficient Inference Acceleration Approach for On-Device Large Language Models","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7137858023","doi":"https://doi.org/10.1609/aaai.v40i31.39789"},"language":null,"primary_location":{"id":"doi:10.1609/aaai.v40i31.39789","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i31.39789","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1609/aaai.v40i31.39789","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129722725","display_name":"Hao Tian","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Hao Tian","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129716089","display_name":"Sheng Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sheng Lu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129706817","display_name":"Fuwen Tian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fuwen Tian","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129670869","display_name":"Guangming Cui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guangming Cui","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129662965","display_name":"Zheng Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng Li","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124064166","display_name":"Xuyun Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xuyun Zhang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129703278","display_name":"Quan Z. Sheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Quan Z. Sheng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5129676742","display_name":"Wanchun Dou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wanchun Dou","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5129722725"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.19553073,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"31","first_page":"25896","last_page":"25904"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.1274999976158142,"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"}},"topics":[{"id":"https://openalex.org/T14347","display_name":"Big Data and Digital Economy","score":0.1274999976158142,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.06849999725818634,"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"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.05959999933838844,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7980999946594238},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7544999718666077},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.7009000182151794},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.6674000024795532},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5149999856948853},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.49070000648498535},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.48980000615119934},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4643000066280365},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.41499999165534973}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8672999739646912},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7980999946594238},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7544999718666077},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.7009000182151794},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.6674000024795532},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5149999856948853},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.49070000648498535},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.48980000615119934},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4643000066280365},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.41499999165534973},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.4059999883174896},{"id":"https://openalex.org/C113174947","wikidata":"https://www.wikidata.org/wiki/Q2859736","display_name":"Tree (set theory)","level":2,"score":0.39469999074935913},{"id":"https://openalex.org/C46743427","wikidata":"https://www.wikidata.org/wiki/Q1341685","display_name":"Inference engine","level":3,"score":0.3889000117778778},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.36469998955726624},{"id":"https://openalex.org/C190290938","wikidata":"https://www.wikidata.org/wiki/Q387015","display_name":"Trie","level":3,"score":0.3441999852657318},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3321000039577484},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.3188999891281128},{"id":"https://openalex.org/C129916263","wikidata":"https://www.wikidata.org/wiki/Q1141183","display_name":"Backward chaining","level":4,"score":0.31279999017715454},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.31119999289512634},{"id":"https://openalex.org/C38556500","wikidata":"https://www.wikidata.org/wiki/Q13404475","display_name":"Cache algorithms","level":4,"score":0.30140000581741333},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2904999852180481},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2892000079154968},{"id":"https://openalex.org/C189783530","wikidata":"https://www.wikidata.org/wiki/Q352090","display_name":"CPU cache","level":3,"score":0.2806999981403351},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.2797999978065491},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.2732999920845032},{"id":"https://openalex.org/C176649486","wikidata":"https://www.wikidata.org/wiki/Q2308807","display_name":"Memory management","level":3,"score":0.263700008392334},{"id":"https://openalex.org/C157170001","wikidata":"https://www.wikidata.org/wiki/Q4781507","display_name":"Applications of artificial intelligence","level":2,"score":0.258899986743927}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1609/aaai.v40i31.39789","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i31.39789","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i31.39789","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i31.39789","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"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],"Models":[2],"(LLMs)":[3],"have":[4],"revolutionized":[5],"intelligent":[6],"interactions,":[7],"enabling":[8,146],"mobile":[9],"applications":[10],"such":[11],"as":[12,26,73,75],"personal":[13],"assistants":[14],"on":[15,51,59,186],"edge":[16,52,111,187],"devices":[17],"for":[18,95,110,133],"local":[19],"execution.":[20],"Speculative":[21],"decoding":[22,151],"(SD)":[23],"has":[24],"emerged":[25],"a":[27,39,90,114,139],"promising":[28],"paradigm":[29],"to":[30,44,127,153,164,176],"accelerate":[31],"LLM":[32],"inference":[33,82,103,199],"without":[34],"compromising":[35],"generation":[36,196],"quality,":[37],"employing":[38],"draft-then-verify":[40],"manner.":[41],"However,":[42],"due":[43],"the":[45,71,155,166,178],"constrained":[46],"computing":[47],"and":[48,69,101,107,131,173,198],"memory":[49,67,106,205],"resources":[50],"devices,":[53],"existing":[54,193],"SD":[55],"works":[56],"heavily":[57],"rely":[58],"an":[60],"auxiliary":[61],"draft":[62],"model":[63],"that":[64,79],"incurs":[65],"additional":[66],"burden":[68],"hinders":[70],"adaptability,":[72],"well":[74],"static":[76],"token":[77,125,141,170],"trees":[78],"yield":[80],"suboptimal":[81],"performance.":[83],"To":[84],"this":[85],"end,":[86],"we":[87],"propose":[88],"DIAA,":[89],"Decoding-efficient":[91],"Inference":[92],"Acceleration":[93],"Approach":[94],"on-device":[96],"LLMs.":[97],"DIAA":[98,137,191],"achieves":[99],"plug-and-play":[100],"model-agnostic":[102],"speedup":[104],"with":[105,143],"computation":[108,159],"efficiency":[109],"devices.":[112],"Specifically,":[113],"pair":[115],"of":[116,169],"lightweight":[117],"look-up":[118],"tables":[119],"(LUTs)":[120],"is":[121,161],"constructed":[122],"by":[123],"Top-K":[124],"sampling":[126],"cache":[128,175],"historical":[129],"tokens":[130],"probabilities":[132],"rapid":[134],"candidate":[135],"drafting.":[136],"integrates":[138],"dynamic":[140],"tree":[142],"prior":[144],"LUTs":[145],"paralleled":[147],"verification,":[148],"updated":[149],"during":[150],"process,":[152],"adapt":[154],"online":[156],"context.":[157],"A":[158],"overlap":[160],"then":[162],"employed":[163],"pipeline":[165],"update":[167],"operations":[168],"tree,":[171],"LUTs,":[172],"KV":[174],"improve":[177],"computational":[179],"efficiency.":[180],"Finally,":[181],"through":[182],"extensive":[183],"experiments":[184],"implemented":[185],"platform":[188],"NVIDIA":[189],"Jetson,":[190],"outperforms":[192],"baselines":[194],"in":[195],"speed":[197],"wall-clock":[200],"time,":[201],"while":[202],"incurring":[203],"minimal":[204],"overhead.":[206]},"counts_by_year":[],"updated_date":"2026-03-18T06:31:55.123368","created_date":"2026-03-18T00:00:00"}