{"id":"https://openalex.org/W4417282553","doi":"https://doi.org/10.1109/pimrc62392.2025.11275395","title":"Collaborative Multi-Device Edge Inference for Vision-Language Models with Speculative Decoding","display_name":"Collaborative Multi-Device Edge Inference for Vision-Language Models with Speculative Decoding","publication_year":2025,"publication_date":"2025-09-01","ids":{"openalex":"https://openalex.org/W4417282553","doi":"https://doi.org/10.1109/pimrc62392.2025.11275395"},"language":"en","primary_location":{"id":"doi:10.1109/pimrc62392.2025.11275395","is_oa":false,"landing_page_url":"https://doi.org/10.1109/pimrc62392.2025.11275395","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 36th International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5023661873","display_name":"Luteng Qiao","orcid":"https://orcid.org/0000-0001-8878-082X"},"institutions":[{"id":"https://openalex.org/I30809798","display_name":"ShanghaiTech University","ror":"https://ror.org/030bhh786","country_code":"CN","type":"education","lineage":["https://openalex.org/I30809798"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Luteng Qiao","raw_affiliation_strings":["ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210","institution_ids":["https://openalex.org/I30809798"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102806605","display_name":"Jiawei Shao","orcid":"https://orcid.org/0000-0001-8836-1430"},"institutions":[{"id":"https://openalex.org/I4210100255","display_name":"Beijing Academy of Artificial Intelligence","ror":"https://ror.org/016a74861","country_code":"CN","type":"other","lineage":["https://openalex.org/I4210100255"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiawei Shao","raw_affiliation_strings":["China Telecom,Institute of Artificial Intelligence (TeleAI)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"China Telecom,Institute of Artificial Intelligence (TeleAI)","institution_ids":["https://openalex.org/I4210100255"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041989556","display_name":"Yong Zhou","orcid":"https://orcid.org/0000-0002-7499-6256"},"institutions":[{"id":"https://openalex.org/I30809798","display_name":"ShanghaiTech University","ror":"https://ror.org/030bhh786","country_code":"CN","type":"education","lineage":["https://openalex.org/I30809798"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yong Zhou","raw_affiliation_strings":["ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210","institution_ids":["https://openalex.org/I30809798"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022499594","display_name":"Yuanming Shi","orcid":"https://orcid.org/0000-0002-1418-7465"},"institutions":[{"id":"https://openalex.org/I30809798","display_name":"ShanghaiTech University","ror":"https://ror.org/030bhh786","country_code":"CN","type":"education","lineage":["https://openalex.org/I30809798"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yuanming Shi","raw_affiliation_strings":["ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ShanghaiTech University,School of Information Science and Technology,Shanghai,China,201210","institution_ids":["https://openalex.org/I30809798"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xuelong Li","orcid":null},"institutions":[{"id":"https://openalex.org/I4210100255","display_name":"Beijing Academy of Artificial Intelligence","ror":"https://ror.org/016a74861","country_code":"CN","type":"other","lineage":["https://openalex.org/I4210100255"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xuelong Li","raw_affiliation_strings":["China Telecom,Institute of Artificial Intelligence (TeleAI)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"China Telecom,Institute of Artificial Intelligence (TeleAI)","institution_ids":["https://openalex.org/I4210100255"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5079052203","display_name":"Khaled B. Letaief","orcid":"https://orcid.org/0000-0003-2519-6401"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Khaled B. Letaief","raw_affiliation_strings":["Hong Kong University of Science and Technology,Department of ECE,Hong Kong,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Hong Kong University of Science and Technology,Department of ECE,Hong Kong,China","institution_ids":["https://openalex.org/I200769079"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"6"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.29910001158714294,"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"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.29910001158714294,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.11749999970197678,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.07509999722242355,"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/inference","display_name":"Inference","score":0.7817000150680542},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.6467999815940857},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.6089000105857849},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.6000999808311462},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.5048999786376953},{"id":"https://openalex.org/keywords/approximate-inference","display_name":"Approximate inference","score":0.4185999929904938},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.3873000144958496}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7907999753952026},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7817000150680542},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.6467999815940857},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.6089000105857849},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.6000999808311462},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.5048999786376953},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.4185999929904938},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.3873000144958496},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3711000084877014},{"id":"https://openalex.org/C158600405","wikidata":"https://www.wikidata.org/wiki/Q5054566","display_name":"Causal inference","level":2,"score":0.36970001459121704},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3190999925136566},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.2897000014781952},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.287200003862381},{"id":"https://openalex.org/C761482","wikidata":"https://www.wikidata.org/wiki/Q118093","display_name":"Transmission (telecommunications)","level":2,"score":0.2858999967575073},{"id":"https://openalex.org/C2778456923","wikidata":"https://www.wikidata.org/wiki/Q5337692","display_name":"Edge computing","level":3,"score":0.2806999981403351},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.2777000069618225},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.27399998903274536},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2653000056743622},{"id":"https://openalex.org/C112313634","wikidata":"https://www.wikidata.org/wiki/Q7886648","display_name":"Complement (music)","level":5,"score":0.26010000705718994}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/pimrc62392.2025.11275395","is_oa":false,"landing_page_url":"https://doi.org/10.1109/pimrc62392.2025.11275395","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 36th International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC)","raw_type":"proceedings-article"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-169663","is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-169663","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"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":"Conference paper"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320309612","display_name":"Natural Science Foundation of Shanghai","ror":null},{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":11,"referenced_works":["https://openalex.org/W3212941463","https://openalex.org/W4306179655","https://openalex.org/W4384080522","https://openalex.org/W4384158407","https://openalex.org/W4400228202","https://openalex.org/W4405361069","https://openalex.org/W4405934565","https://openalex.org/W4406856790","https://openalex.org/W4407449947","https://openalex.org/W4414538737","https://openalex.org/W4414539726"],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"have":[4],"demonstrated":[5],"remarkable":[6],"success":[7],"across":[8,47],"various":[9],"domains.":[10],"However,":[11,50],"the":[12,25,32,40,51,59,67,90,143,148,155,161,176,189],"substantial":[13],"computational":[14],"and":[15],"memory":[16],"demands":[17],"pose":[18],"significant":[19],"challenges":[20],"for":[21],"deploying":[22],"LLMs":[23,43,46],"at":[24,89,182],"network":[26,91],"edge.":[27],"To":[28,73],"address":[29],"this":[30,101],"issue,":[31],"existing":[33],"studies":[34],"mainly":[35],"focus":[36],"on":[37],"either":[38],"reducing":[39],"size":[41],"of":[42,69,145,164,178],"or":[44],"distributing":[45],"multiple":[48],"devices.":[49],"former":[52],"approach":[53],"suffers":[54],"from":[55,140],"performance":[56],"degradation,":[57],"while":[58],"latter":[60],"incurs":[61],"high":[62],"communication":[63],"cost":[64],"due":[65],"to":[66,83,111,124,147,174,198],"transmission":[68],"high-dimensional":[70],"intermediate":[71],"features.":[72],"tackle":[74],"these":[75],"issues,":[76],"we":[77,159],"propose":[78],"a":[79,97,135,169,193],"novel":[80],"collaborative":[81],"framework":[82],"support":[84],"vision-language":[85],"model":[86],"(VLM)":[87],"inference":[88,156,165],"edge,":[92],"expanding":[93],"speculative":[94,141],"decoding":[95],"into":[96],"multi-device":[98],"scenario.":[99],"In":[100],"framework,":[102],"each":[103,183],"edge":[104,126],"device":[105],"first":[106],"utilizes":[107],"its":[108],"small":[109],"VLM":[110,150],"generate":[112],"draft":[113,179],"tokens":[114,120,180],"in":[115,132],"an":[116,125],"auto-regressive":[117],"manner.":[118],"These":[119],"are":[121,130],"then":[122],"transmitted":[123],"server,":[127],"where":[128],"they":[129],"corrected":[131],"parallel":[133],"by":[134,167],"large":[136,149],"VLM.":[137],"By":[138],"benefiting":[139],"decoding,":[142],"number":[144,177],"calls":[146],"is":[151],"reduced":[152],"without":[153],"degrading":[154],"performance.":[157],"Furthermore,":[158],"minimize":[160],"average":[162,195],"latency":[163,196],"tasks":[166],"developing":[168],"deep":[170],"reinforcement":[171],"learning":[172],"algorithm":[173,191],"optimize":[175],"generated":[181],"iteration.":[184],"Simulation":[185],"results":[186],"confirm":[187],"that":[188],"proposed":[190],"achieves":[192],"lower":[194],"compared":[197],"other":[199],"baselines.":[200]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-12-12T00:00:00"}