{"id":"https://openalex.org/W7161677174","doi":"https://doi.org/10.1109/tccn.2026.3694872","title":"Efficient Edge Deployment of Mixture-of-Experts Models With Hybrid Expert Quantization","display_name":"Efficient Edge Deployment of Mixture-of-Experts Models With Hybrid Expert Quantization","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7161677174","doi":"https://doi.org/10.1109/tccn.2026.3694872"},"language":null,"primary_location":{"id":"doi:10.1109/tccn.2026.3694872","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tccn.2026.3694872","pdf_url":null,"source":{"id":"https://openalex.org/S2484188435","display_name":"IEEE Transactions on Cognitive Communications and Networking","issn_l":"2332-7731","issn":["2332-7731","2372-2045"],"is_oa":false,"is_in_doaj":false,"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 Transactions on Cognitive Communications and Networking","raw_type":"journal-article"},"type":"article","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/A5100338120","display_name":"Tao Liu","orcid":"https://orcid.org/0000-0002-8201-5307"},"institutions":[{"id":"https://openalex.org/I141591182","display_name":"University of Aizu","ror":"https://ror.org/02pg0e883","country_code":"JP","type":"education","lineage":["https://openalex.org/I141591182"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Tao Liu","raw_affiliation_strings":["School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan"],"raw_orcid":"https://orcid.org/0000-0001-8261-9277","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan","institution_ids":["https://openalex.org/I141591182"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136499294","display_name":"Qipeng Xie","orcid":"https://orcid.org/0000-0002-5500-0249"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qipeng Xie","raw_affiliation_strings":["The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-5500-0249","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5063664617","display_name":"Zhaoyang Han","orcid":"https://orcid.org/0000-0002-1511-361X"},"institutions":[{"id":"https://openalex.org/I167027274","display_name":"Nanjing Forestry University","ror":"https://ror.org/03m96p165","country_code":"CN","type":"education","lineage":["https://openalex.org/I167027274"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhaoyang Han","raw_affiliation_strings":["College of Information Science and Technology and Artificial Intelligence, Nanjing Forestry University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0002-1416-4677","affiliations":[{"raw_affiliation_string":"College of Information Science and Technology and Artificial Intelligence, Nanjing Forestry University, Nanjing, China","institution_ids":["https://openalex.org/I167027274"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136457491","display_name":"Weizheng Wang","orcid":"https://orcid.org/0000-0002-5879-585X"},"institutions":[{"id":"https://openalex.org/I37987034","display_name":"Guangzhou University","ror":"https://ror.org/05ar8rn06","country_code":"CN","type":"education","lineage":["https://openalex.org/I37987034"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Weizheng Wang","raw_affiliation_strings":["Cyberspace Institute Advanced Technology, Guangzhou University, Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-5879-585X","affiliations":[{"raw_affiliation_string":"Cyberspace Institute Advanced Technology, Guangzhou University, Guangzhou, China","institution_ids":["https://openalex.org/I37987034"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010520102","display_name":"Shengli Pan","orcid":null},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shengli Pan","raw_affiliation_strings":["School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0002-1904-6557","affiliations":[{"raw_affiliation_string":"School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing, China","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082442712","display_name":"Xiang Zhang","orcid":"https://orcid.org/0000-0003-0413-6135"},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiang Zhang","raw_affiliation_strings":["School of Cyber Security, Tianjin University, Tianjin, China"],"raw_orcid":"https://orcid.org/0000-0003-0413-6135","affiliations":[{"raw_affiliation_string":"School of Cyber Security, Tianjin University, Tianjin, China","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136479604","display_name":"Chunhua Su","orcid":"https://orcid.org/0000-0002-6461-9684"},"institutions":[{"id":"https://openalex.org/I141591182","display_name":"University of Aizu","ror":"https://ror.org/02pg0e883","country_code":"JP","type":"education","lineage":["https://openalex.org/I141591182"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Chunhua Su","raw_affiliation_strings":["School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan"],"raw_orcid":"https://orcid.org/0000-0002-6461-9684","affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan","institution_ids":["https://openalex.org/I141591182"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5136489200","display_name":"Kaishun Wu","orcid":"https://orcid.org/0000-0003-2216-0737"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kaishun Wu","raw_affiliation_strings":["The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-2216-0737","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.60960207,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"12","issue":null,"first_page":"10164","last_page":"10178"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.25870001316070557,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.25870001316070557,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.1152999997138977,"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/T11901","display_name":"Bayesian Methods and Mixture Models","score":0.06419999897480011,"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/software-deployment","display_name":"Software deployment","score":0.5275999903678894},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.5078999996185303},{"id":"https://openalex.org/keywords/enhanced-data-rates-for-gsm-evolution","display_name":"Enhanced Data Rates for GSM Evolution","score":0.3589000105857849},{"id":"https://openalex.org/keywords/vector-quantization","display_name":"Vector quantization","score":0.2849999964237213},{"id":"https://openalex.org/keywords/signal-processing","display_name":"Signal processing","score":0.28299999237060547},{"id":"https://openalex.org/keywords/algorithm-design","display_name":"Algorithm design","score":0.26330000162124634}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8086000084877014},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5275999903678894},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.5078999996185303},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4088999927043915},{"id":"https://openalex.org/C162307627","wikidata":"https://www.wikidata.org/wiki/Q204833","display_name":"Enhanced Data Rates for GSM Evolution","level":2,"score":0.3589000105857849},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.32339999079704285},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3061000108718872},{"id":"https://openalex.org/C199833920","wikidata":"https://www.wikidata.org/wiki/Q612536","display_name":"Vector quantization","level":2,"score":0.2849999964237213},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.28299999237060547},{"id":"https://openalex.org/C106516650","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm design","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.26159998774528503},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2578999996185303},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.25690001249313354},{"id":"https://openalex.org/C158457486","wikidata":"https://www.wikidata.org/wiki/Q17104301","display_name":"Multidimensional systems","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tccn.2026.3694872","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tccn.2026.3694872","pdf_url":null,"source":{"id":"https://openalex.org/S2484188435","display_name":"IEEE Transactions on Cognitive Communications and Networking","issn_l":"2332-7731","issn":["2332-7731","2372-2045"],"is_oa":false,"is_in_doaj":false,"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 Transactions on Cognitive Communications and Networking","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.521940290927887,"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy"}],"awards":[{"id":"https://openalex.org/G1809975003","display_name":null,"funder_award_id":"62271066","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":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Deploying":[0],"large":[1],"language":[2],"models":[3,71],"(LLMs)":[4],"directly":[5],"at":[6,34],"the":[7,29,35,42,45,76,83,111,141,155,164,177],"edge":[8,59,63,87,102],"offers":[9],"significant":[10],"benefits":[11],"in":[12,56],"reducing":[13],"inference":[14,137],"latency":[15],"and":[16,128,135],"enhancing":[17],"privacy":[18],"protection.":[19],"However,":[20],"this":[21,38],"approach":[22,189],"presents":[23],"critical":[24],"technical":[25],"challenges":[26],"due":[27],"to":[28,106,132],"limited":[30],"hardware":[31],"resources":[32],"available":[33],"edge.":[36],"In":[37],"paper,":[39],"we":[40,93,117,146,162],"explore":[41],"deployment":[43],"of":[44,79,86,113,159],"Mixture-of-Experts":[46],"(MoE)":[47],"model,":[48],"a":[49,57,95,119,148,168],"popular":[50],"LLM":[51],"architecture,":[52],"using":[53,167],"weight":[54],"quantization":[55,98,127],"typical":[58,158],"environment":[60],"with":[61],"multiple":[62],"servers.":[64],"Existing":[65],"methods":[66],"that":[67,123,171,187],"uniformly":[68],"quantize":[69],"MoE":[70],"do":[72],"not":[73],"account":[74],"for":[75],"varying":[77],"importance":[78],"different":[80,107],"experts":[81,105],"or":[82],"dynamic":[84],"nature":[85],"requests.":[88,115],"To":[89,139,153],"address":[90],"these":[91],"issues,":[92],"propose":[94],"hybrid":[96],"expert":[97,126],"strategy,":[99],"where":[100],"each":[101],"server":[103],"quantizes":[104],"bit-widths":[108],"based":[109],"on":[110,183],"preferences":[112],"received":[114],"Furthermore,":[116],"implement":[118],"comprehensive":[120],"system":[121,169],"optimization":[122,144],"jointly":[124],"considers":[125],"request":[129],"assignment,":[130],"aiming":[131],"achieve":[133],"high-precision":[134],"low-latency":[136],"services.":[138],"manage":[140],"complex,":[142],"coupled":[143],"challenges,":[145],"design":[147],"reinforcement":[149],"learning":[150,178],"(RL)-based":[151],"algorithm.":[152],"overcome":[154],"slow":[156],"convergence":[157],"traditional":[160],"RL,":[161],"pre-train":[163],"policy":[165],"network":[166],"simulator":[170],"generates":[172],"heuristic":[173],"solutions,":[174],"thus":[175],"accelerating":[176],"process.":[179],"Extensive":[180],"experimental":[181],"results":[182],"real":[184],"datasets":[185],"demonstrate":[186],"our":[188],"significantly":[190],"outperforms":[191],"existing":[192],"methods.":[193]},"counts_by_year":[],"updated_date":"2026-07-21T05:50:15.549392","created_date":"2026-05-20T00:00:00"}