{"id":"https://openalex.org/W7155607794","doi":"https://doi.org/10.1016/j.neunet.2026.109014","title":"Optimizing speculative decoding via dynamic multi-path and dual-stream networks","display_name":"Optimizing speculative decoding via dynamic multi-path and dual-stream networks","publication_year":2026,"publication_date":"2026-04-25","ids":{"openalex":"https://openalex.org/W7155607794","doi":"https://doi.org/10.1016/j.neunet.2026.109014","pmid":"https://pubmed.ncbi.nlm.nih.gov/42068634"},"language":"en","primary_location":{"id":"doi:10.1016/j.neunet.2026.109014","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.neunet.2026.109014","pdf_url":null,"source":{"id":"https://openalex.org/S123019304","display_name":"Neural Networks","issn_l":"0893-6080","issn":["0893-6080","1879-2782"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural Networks","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.1016/j.neunet.2026.109014","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134600813","display_name":"Yong Yang","orcid":null},"institutions":[{"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":true,"raw_author_name":"Yong Yang","raw_affiliation_strings":["School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China; Pengcheng Laboratory, Shenzhen, China. Electronic address: 934686440@qq.com"],"raw_orcid":"https://orcid.org/0000-0002-3309-4643","affiliations":[{"raw_affiliation_string":"School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China; Pengcheng Laboratory, Shenzhen, China. Electronic address: 934686440@qq.com","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101702056","display_name":"Ting Yang","orcid":"https://orcid.org/0000-0001-8586-1675"},"institutions":[{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ting Ting Yang","raw_affiliation_strings":["Pengcheng Laboratory, Shenzhen, China. Electronic address: yangtt@pcl.ac.cn"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pengcheng Laboratory, Shenzhen, China. Electronic address: yangtt@pcl.ac.cn","institution_ids":["https://openalex.org/I4210136793"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126730561","display_name":"Shao Shuai Gao","orcid":null},"institutions":[{"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":"Shao Shuai Gao","raw_affiliation_strings":["School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China","institution_ids":["https://openalex.org/I4210165038"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134611490","display_name":"Jia Hong Ning","orcid":null},"institutions":[{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jia Hong Ning","raw_affiliation_strings":["Pengcheng Laboratory, Shenzhen, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pengcheng Laboratory, Shenzhen, China","institution_ids":["https://openalex.org/I4210136793"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5134586990","display_name":"Ling Zheng Kong","orcid":null},"institutions":[{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ling Zheng Kong","raw_affiliation_strings":["Pengcheng Laboratory, Shenzhen, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Pengcheng Laboratory, Shenzhen, China","institution_ids":["https://openalex.org/I4210136793"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5134600813"],"corresponding_institution_ids":["https://openalex.org/I4210165038"],"apc_list":{"value":3350,"currency":"USD","value_usd":3350},"apc_paid":{"value":3350,"currency":"USD","value_usd":3350},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.48447739,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"202","issue":null,"first_page":"109014","last_page":"109014"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.20990000665187836,"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":0.20990000665187836,"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":0.13830000162124634,"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/T13910","display_name":"Computational and Text Analysis Methods","score":0.08669999986886978,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7419999837875366},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.7279999852180481},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.6852999925613403},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.5730999708175659},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.41440001130104065},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.3921000063419342},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.320499986410141}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8766000270843506},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7419999837875366},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7279999852180481},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.6852999925613403},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.5730999708175659},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.41440001130104065},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.3921000063419342},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3393000066280365},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.320499986410141},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3158000111579895},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.31200000643730164},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.2854999899864197},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2809000015258789},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.2782000005245209},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.26930001378059387},{"id":"https://openalex.org/C89198739","wikidata":"https://www.wikidata.org/wiki/Q3079880","display_name":"Data stream mining","level":2,"score":0.26030001044273376},{"id":"https://openalex.org/C186644900","wikidata":"https://www.wikidata.org/wiki/Q194152","display_name":"Parsing","level":2,"score":0.2531999945640564},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25119999051094055}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1016/j.neunet.2026.109014","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.neunet.2026.109014","pdf_url":null,"source":{"id":"https://openalex.org/S123019304","display_name":"Neural Networks","issn_l":"0893-6080","issn":["0893-6080","1879-2782"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural Networks","raw_type":"journal-article"},{"id":"pmid:42068634","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/42068634","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural networks : the official journal of the International Neural Network Society","raw_type":null}],"best_oa_location":{"id":"doi:10.1016/j.neunet.2026.109014","is_oa":true,"landing_page_url":"https://doi.org/10.1016/j.neunet.2026.109014","pdf_url":null,"source":{"id":"https://openalex.org/S123019304","display_name":"Neural Networks","issn_l":"0893-6080","issn":["0893-6080","1879-2782"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320990","host_organization_name":"Elsevier BV","host_organization_lineage":["https://openalex.org/P4310320990"],"host_organization_lineage_names":["Elsevier BV"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural Networks","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320318558","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W4389523983","https://openalex.org/W4402671766","https://openalex.org/W4402683901","https://openalex.org/W4412888552","https://openalex.org/W7133225605","https://openalex.org/W7133239970","https://openalex.org/W7133245307"],"related_works":[],"abstract_inverted_index":{"The":[0,105],"Transformer":[1],"architecture":[2,91],"widely":[3],"adopted":[4],"in":[5,114,188],"the":[6,18,55,59,66,69,80,90,119,135,140,155,168,177,181],"large":[7],"language":[8],"models":[9,44],"(LLMs)":[10],"suffers":[11],"from":[12],"limited":[13],"inference":[14,39,169,190],"efficiency":[15],"due":[16],"to":[17,36,46,117,133,150,157],"inherently":[19],"sequential":[20],"nature":[21],"of":[22,73,107,121],"autoregressive":[23],"token":[24,148],"generation.":[25],"To":[26,78],"address":[27],"this":[28,82],"issue,":[29],"speculative":[30,43],"decoding":[31,128],"(SD)":[32],"has":[33],"been":[34],"proposed":[35,176],"accelerate":[37],"LLM":[38],"by":[40,54,65,76,139],"employing":[41],"small":[42],"(SSMs)":[45],"generate":[47],"candidate":[48,122,147],"tokens":[49,74,160],"that":[50,97,174],"are":[51,111],"subsequently":[52],"verified":[53],"target":[56],"LLM.":[57],"However,":[58],"SD":[60,183],"methods":[61],"is":[62,131],"often":[63],"constrained":[64],"key":[67],"challenges:":[68],"low":[70],"acceptance":[71],"rate":[72],"predicted":[75],"SSMs.":[77],"overcome":[79],"limitation,":[81],"paper":[83],"proposes":[84],"a":[85,125,162],"Dual-Stream":[86],"Network":[87],"Architecture":[88],"(DSNA),":[89],"introduces":[92],"two":[93,109],"parallel":[94],"processing":[95],"streams":[96,110],"simultaneously":[98],"model":[99,156],"word":[100],"sequences":[101],"and":[102,192],"feature":[103],"sequences.":[104],"outputs":[106],"these":[108],"progressively":[112],"fused":[113],"subsequent":[115],"stages":[116],"enhance":[118],"quality":[120],"predictions.":[123],"Furthermore,":[124],"dynamic":[126],"multi-path":[127],"(DMPD)":[129],"mechanism":[130,144],"introduced":[132],"leverage":[134],"enriched":[136],"representations":[137],"produced":[138],"dual-stream":[141],"architecture.":[142],"This":[143],"allows":[145],"multiple":[146,159,196],"paths":[149],"be":[151],"evaluated":[152],"simultaneously,":[153],"enabling":[154],"accept":[158],"within":[161],"single":[163],"forward":[164],"propagation":[165],"step":[166],"during":[167],"process.":[170],"Extensive":[171],"experiments":[172],"show":[173],"our":[175],"method":[178],"consistently":[179],"outperforms":[180],"state-of-the-art":[182],"approaches,":[184],"achieving":[185],"significant":[186],"improvements":[187],"both":[189],"throughput":[191],"generation":[193],"accuracy":[194],"across":[195],"benchmarks.":[197]},"counts_by_year":[],"updated_date":"2026-06-23T06:36:01.041984","created_date":"2026-04-26T00:00:00"}