{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,28]],"date-time":"2026-10-28T00:03:21Z","timestamp":1793145801888,"version":"4.5.0"},"reference-count":56,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100002367","name":"Key Deployment Program of Chinese Academy of Sciences, China","doi-asserted-by":"publisher","award":["KGFZD-145-25-39"],"award-info":[{"award-number":["KGFZD-145-25-39"]}],"id":[{"id":"10.13039\/501100002367","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Strategic Priority Research Program of Chinese Academy of Sciences, China","award":["E1XA310103"],"award-info":[{"award-number":["E1XA310103"]}],"id":[{"id":"https:\/\/ror.org\/034t30j35","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. on Image Process."],"published-print":{"date-parts":[[2025]]},"DOI":"10.1109\/tip.2025.3626887","type":"journal-article","created":{"date-parts":[[2025,11,4]],"date-time":"2025-11-04T18:38:24Z","timestamp":1762281504000},"page":"7209-7221","source":"Crossref","is-referenced-by-count":22,"title":["ViMoE: An Empirical Study of Designing Vision Mixture-of-Experts"],"prefix":"10.1109","volume":"34","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1636-463X","authenticated-orcid":false,"given":"Xumeng","family":"Han","sequence":"first","affiliation":[{"name":"University of Chinese Academy of Sciences (UCAS), Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6916-3009","authenticated-orcid":false,"given":"Longhui","family":"Wei","sequence":"additional","affiliation":[{"name":"Huawei Inc., Shenzhen, Guangdong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhiyang","family":"Dou","sequence":"additional","affiliation":[{"name":"University of Chinese Academy of Sciences (UCAS), Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yingfei","family":"Sun","sequence":"additional","affiliation":[{"name":"University of Chinese Academy of Sciences (UCAS), Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9970-5152","authenticated-orcid":false,"given":"Zhenjun","family":"Han","sequence":"additional","affiliation":[{"name":"University of Chinese Academy of Sciences (UCAS), Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7252-5047","authenticated-orcid":false,"given":"Qi","family":"Tian","sequence":"additional","affiliation":[{"name":"Huawei Inc., Shenzhen, Guangdong, 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":"Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context","author":"Team","year":"2024","journal-title":"arXiv:2403.05530"},{"key":"ref3","article-title":"Qwen2 technical report","volume-title":"arXiv:2407.10671","author":"Yang","year":"2024"},{"key":"ref4","article-title":"The llama 3 herd of models","author":"Grattafiori","year":"2024","journal-title":"arXiv:2407.21783"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.1.79"},{"key":"ref6","article-title":"Learning factored representations in a deep mixture of experts","author":"Eigen","year":"2013","journal-title":"arXiv:1312.4314"},{"key":"ref7","article-title":"Mixtral of experts","author":"Jiang","year":"2024","journal-title":"arXiv:2401.04088"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.70"},{"key":"ref9","article-title":"OpenMoE: An early effort on open mixture-of-experts language models","author":"Xue","year":"2024","journal-title":"arXiv:2402.01739"},{"key":"ref10","article-title":"MoE-LLaVA: Mixture of experts for large vision-language models","author":"Lin","year":"2024","journal-title":"arXiv:2401.15947"},{"key":"ref11","article-title":"OLMoE: Open mixture-of-experts language models","author":"Muennighoff","year":"2024","journal-title":"arXiv:2409.02060"},{"key":"ref12","article-title":"Aria: An open multimodal native mixture-of-experts model","author":"Li","year":"2024","journal-title":"arXiv:2410.05993"},{"key":"ref13","first-page":"611","article-title":"An image is worth 16\u00d716 words: Transformers for image recognition at scale","volume-title":"Proc. ICLR","author":"Dosovitskiy"},{"key":"ref14","first-page":"8583","article-title":"Scaling vision with sparse mixture of experts","volume-title":"Proc. NIPS","author":"Riquelme"},{"key":"ref15","article-title":"Mobile V-MoEs: Scaling down vision transformers via sparse mixture-of-experts","author":"Daxberger","year":"2023","journal-title":"arXiv:2309.04354"},{"key":"ref16","article-title":"Task-customized masked AutoEncoder via mixture of cluster-conditional experts","author":"Liu","year":"2024","journal-title":"arXiv:2402.05382"},{"key":"ref17","article-title":"From sparse to soft mixtures of experts","author":"Puigcerver","year":"2023","journal-title":"arXiv:2308.00951"},{"key":"ref18","article-title":"Residual mixture of experts","author":"Wu","year":"2022","journal-title":"arXiv:2204.09636"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i8.20858"},{"key":"ref20","article-title":"DINOv2: Learning robust visual features without supervision","author":"Oquab","year":"2023","journal-title":"arXiv:2304.07193"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref22","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Touvron"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20053-3_30"},{"key":"ref24","article-title":"BEiT: BERT pre-training of image transformers","author":"Bao","year":"2021","journal-title":"arXiv:2106.08254"},{"key":"ref25","article-title":"IBOT: Image BERT pre-training with online tokenizer","author":"Zhou","year":"2021","journal-title":"arXiv:2111.07832"},{"key":"ref26","article-title":"DINO: DETR with improved DeNoising anchor boxes for end-to-end object detection","author":"Zhang","year":"2022","journal-title":"arXiv:2203.03605"},{"key":"ref27","article-title":"An empirical study of training self-supervised visual transformers","author":"Xinlei","year":"2021","journal-title":"arXiv:2104.02057"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01553"},{"key":"ref29","first-page":"6265","article-title":"BASE layers: Simplifying training of large, sparse models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Lewis"},{"key":"ref30","first-page":"7103","article-title":"Mixture-of-experts with expert choice routing","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Zhou"},{"key":"ref31","first-page":"18332","article-title":"DeepSpeed-MoE: Advancing mixture-of-experts inference and training to power next-generation AI scale","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Rajbhandari"},{"key":"ref32","article-title":"Mixture of cluster-conditional LoRA experts for vision-language instruction tuning","author":"Gou","year":"2023","journal-title":"arXiv:2312.12379"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.890"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3103132"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2012.2200299"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-012-9338-y"},{"key":"ref37","first-page":"269","article-title":"Tutel: Adaptive mixture-of-experts at scale","volume-title":"Proc. Mach. Learn. Syst.","author":"Hwang"},{"key":"ref38","first-page":"29335","article-title":"DSelect-k: Differentiable selection in the mixture of experts with applications to multi-task learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Hazimeh"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2021.3117837"},{"issue":"120","key":"ref40","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":"ref41","article-title":"GShard: Scaling giant models with conditional computation and automatic sharding","author":"Lepikhin","year":"2020","journal-title":"arXiv:2006.16668"},{"key":"ref42","first-page":"5547","article-title":"GLaM: Efficient scaling of language models with mixture-of-experts","volume-title":"Proc. Int. Conf. Mach. Learn.","volume":"162","author":"Du"},{"key":"ref43","first-page":"28441","article-title":"M3vit: Mixture-of-experts vision transformer for efficient multi-task learning with model-accelerator co-design","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"35","author":"Fan"},{"key":"ref44","first-page":"6074","article-title":"Patch-level routing in mixture-of-experts is provably sample-efficient for convolutional neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Chowdhury"},{"key":"ref45","article-title":"Scaling diffusion transformers to 16 billion parameters","author":"Fei","year":"2024","journal-title":"arXiv:2407.11633"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01591"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.2.181"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref50","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","author":"Shazeer","year":"2017","journal-title":"arXiv:1701.06538"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.02153"},{"key":"ref52","article-title":"ST-MoE: Designing stable and transferable sparse expert models","author":"Zoph","year":"2022","journal-title":"arXiv:2202.08906"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.544"},{"key":"ref54","first-page":"4061","article-title":"Decoupled weight decay regularization","volume-title":"Proc. ICLR","author":"Loshchilov"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01228-1_26"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.683"}],"container-title":["IEEE Transactions on Image Processing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/83\/10795784\/11226953.pdf?arnumber=11226953","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,17]],"date-time":"2025-11-17T18:43:36Z","timestamp":1763405016000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11226953\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025]]},"references-count":56,"URL":"https:\/\/doi.org\/10.1109\/tip.2025.3626887","relation":{},"ISSN":["1057-7149","1941-0042"],"issn-type":[{"value":"1057-7149","type":"print"},{"value":"1941-0042","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025]]}}}