{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T05:07:19Z","timestamp":1784610439623,"version":"3.55.0"},"reference-count":55,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"JSPS Grant-in-Aid for Scientific Research Kiban","award":["23K11103"],"award-info":[{"award-number":["23K11103"]}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["62271066"],"award-info":[{"award-number":["62271066"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Cogn. Commun. Netw."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/tccn.2026.3694872","type":"journal-article","created":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T19:50:58Z","timestamp":1779220258000},"page":"10164-10178","source":"Crossref","is-referenced-by-count":0,"title":["Efficient Edge Deployment of Mixture-of-Experts Models With Hybrid Expert Quantization"],"prefix":"10.1109","volume":"12","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8261-9277","authenticated-orcid":false,"given":"Tao","family":"Liu","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5500-0249","authenticated-orcid":false,"given":"Qipeng","family":"Xie","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1416-4677","authenticated-orcid":false,"given":"Zhaoyang","family":"Han","sequence":"additional","affiliation":[{"name":"College of Information Science and Technology and Artificial Intelligence, Nanjing Forestry University, Nanjing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5879-585X","authenticated-orcid":false,"given":"Weizheng","family":"Wang","sequence":"additional","affiliation":[{"name":"Cyberspace Institute Advanced Technology, Guangzhou University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1904-6557","authenticated-orcid":false,"given":"Shengli","family":"Pan","sequence":"additional","affiliation":[{"name":"School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0413-6135","authenticated-orcid":false,"given":"Xiang","family":"Zhang","sequence":"additional","affiliation":[{"name":"School of Cyber Security, Tianjin University, Tianjin, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6461-9684","authenticated-orcid":false,"given":"Chunhua","family":"Su","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2216-0737","authenticated-orcid":false,"given":"Kaishun","family":"Wu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"GPT-4 technical report","volume-title":"arXiv:2303.08774","author":"Achiam","year":"2023"},{"key":"ref2","article-title":"PrivateLoRA for efficient privacy preserving LLM","author":"Wang","year":"2023","journal-title":"arXiv:2311.14030"},{"key":"ref3","article-title":"EdgeMoE: Empowering sparse large language models on mobile devices","author":"Yi","year":"2023","journal-title":"arXiv:2308.14352"},{"key":"ref4","article-title":"SwapMoE: Serving off-the-shelf MoE-based large language models with tunable memory budget","author":"Kong","year":"2023","journal-title":"arXiv:2308.15030"},{"key":"ref5","article-title":"Fiddler: CPU\u2013GPU orchestration for fast inference of mixture-of-experts models","author":"Kamahori","year":"2024","journal-title":"arXiv:2402.07033"},{"key":"ref6","article-title":"MoE-infinity: Efficient MoE inference on personal machines with sparsity-aware expert cache","author":"Xue","year":"2024","journal-title":"arXiv:2401.14361"},{"key":"ref7","article-title":"Fast inference of mixture-of-experts language models with offloading","author":"Eliseev","year":"2023","journal-title":"arXiv:2312.17238"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.findings-emnlp.304"},{"key":"ref9","article-title":"Task-specific expert pruning for sparse mixture-of-experts","author":"Chen","year":"2022","journal-title":"arXiv:2206.00277"},{"key":"ref10","first-page":"439","article-title":"Qmoe: Sub-1-bit compression of trillion parameter models","volume":"6","author":"Frantar","year":"2024","journal-title":"Proc. Mach. Learn. Syst."},{"key":"ref11","article-title":"Mixture of quantized experts (MoQE): Complementary effect of low-bit quantization and robustness","author":"Kim","year":"2023","journal-title":"arXiv:2310.02410"},{"key":"ref12","article-title":"Optq: Accurate quantization for generative pre-trained transformers","volume-title":"Proc. 11th Int. Conf. Learn. Represent.","author":"Frantar"},{"key":"ref13","first-page":"87","article-title":"AWQ: Activation-aware weight quantization for on-device llm compression and acceleration","volume-title":"Proc. Mach. Learn. Syst.","volume":"6","author":"Lin"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i12.29237"},{"key":"ref15","article-title":"LeanQuant: Accurate and scalable large language model quantization with loss-error-aware grid","volume-title":"arXiv:2407.10032","author":"Zhang","year":"2024"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.naacl-long.543"},{"key":"ref17","first-page":"27168","article-title":"ZeroQuant: Efficient and affordable post-training quantization for large-scale transformers","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"He"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2198"},{"key":"ref19","first-page":"38087","article-title":"SmoothQuant: Accurate and efficient post-training quantization for large language models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Xiao"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i17.29860"},{"key":"ref21","article-title":"PowerInfer: Fast large language model serving with a consumer-grade GPU","author":"Song","year":"2023","journal-title":"arXiv:2312.12456"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2181"},{"key":"ref23","article-title":"Toward cost-efficient serving of mixture-of-experts with asynchrony","author":"Wang","year":"2025","journal-title":"arXiv:2505.08944"},{"key":"ref24","article-title":"QoS-efficient serving of multiple mixture-of-expert LLMs using partial runtime reconfiguration","volume-title":"arXiv:2505.06481","author":"Imani","year":"2025"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/TON.2026.3669011"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2026.3656318"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.1.79"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.2.181"},{"key":"ref29","article-title":"Mixtral of experts","author":"Jiang","year":"2024","journal-title":"arXiv:2401.04088"},{"key":"ref30","volume-title":"Qwen2moe","year":"2023"},{"issue":"120","key":"ref31","first-page":"1","article-title":"Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity","volume":"23","author":"Fedus","year":"2022","journal-title":"J. Mach. Learn. Res."},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.70"},{"key":"ref33","article-title":"KIVI: A tuning-free asymmetric 2bit quantization for KV cache","author":"Liu","year":"2024","journal-title":"arXiv:2402.02750"},{"key":"ref34","article-title":"No token left behind: Reliable KV cache compression via importance-aware mixed precision quantization","author":"Yang","year":"2024","journal-title":"arXiv:2402.18096"},{"key":"ref35","volume-title":"Qwen-Based MoE Model","year":"2024"},{"key":"ref36","article-title":"Measuring massive multitask language understanding","volume-title":"arXiv:2009.03300","author":"Hendrycks","year":"2020"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1817"},{"key":"ref38","first-page":"14234","article-title":"Merge, then compress: Demystify efficient smoe with hints from its routing policy","volume-title":"Proc. Int. Conf. Learn. Represent.","volume":"2024","author":"Li"},{"key":"ref39","volume-title":"Llama Moe","year":"2024"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1102"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-5409"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P18-1205"},{"key":"ref44","volume-title":"Yelp Dataset","year":"2024"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2025.3596173"},{"key":"ref46","article-title":"Proximal policy optimization algorithms","author":"Schulman","year":"2017","journal-title":"arXiv:1707.06347"},{"key":"ref47","article-title":"High-dimensional continuous control using generalized advantage estimation","author":"Schulman","year":"2016","journal-title":"ICLR"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/MIC.2024.3460908"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1109\/MIC.2023.3284693"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9781139014830"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.3758\/BF03206555"},{"key":"ref52","volume-title":"Half-Quadratic Quantization of Large Machine Learning Models","author":"Badri","year":"2023"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/VTC2023-Fall60731.2023.10333824"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/TON.2025.3585359"},{"key":"ref55","article-title":"UserCentrix: An agentic memory-augmented AI framework for smart spaces","author":"Saleh","year":"2025","journal-title":"arXiv:2505.00472"}],"container-title":["IEEE Transactions on Cognitive Communications and Networking"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6687307\/11304002\/11527015.pdf?arnumber=11527015","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,21]],"date-time":"2026-07-21T04:55:37Z","timestamp":1784609737000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11527015\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":55,"URL":"https:\/\/doi.org\/10.1109\/tccn.2026.3694872","relation":{},"ISSN":["2332-7731","2372-2045"],"issn-type":[{"value":"2332-7731","type":"electronic"},{"value":"2372-2045","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}