{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T10:07:35Z","timestamp":1777889255534,"version":"3.51.4"},"reference-count":44,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["92467204,62472249"],"award-info":[{"award-number":["92467204,62472249"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01436","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"15481-15490","source":"Crossref","is-referenced-by-count":0,"title":["DICE: Staleness-Centric Optimizations for Parallel Diffusion MoE Inference"],"prefix":"10.1109","author":[{"given":"Jiajun","family":"Luo","sequence":"first","affiliation":[{"name":"Shenzhen International Graduate School, Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lizhuo","family":"Luo","sequence":"additional","affiliation":[{"name":"Southern University of Science and Technology 3The Chinese University of Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jianru","family":"Xu","sequence":"additional","affiliation":[{"name":"Southern University of Science and Technology 3The Chinese University of Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jiajun","family":"Song","sequence":"additional","affiliation":[{"name":"Shenzhen International Graduate School, Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rongwei","family":"Lu","sequence":"additional","affiliation":[{"name":"Shenzhen International Graduate School, Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chen","family":"Tang","sequence":"additional","affiliation":[{"name":"Southern University of Science and Technology 3The Chinese University of Hong Kong"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhi","family":"Wang","sequence":"additional","affiliation":[{"name":"Shenzhen International Graduate School, Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52734.2025.02636"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3015"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref4","first-page":"8780","article-title":"Diffusion models beat gans on image synthesis","volume-title":"Advances in neural information processing systems","volume":"34","author":"Dhariwal","year":"2021"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.emnlp-main.1486"},{"key":"ref6","author":"Fei","year":"2024","journal-title":"feizc\/DiT-MoE"},{"key":"ref7","article-title":"Scaling diffusion transformers to 16 billion parameters","author":"Fei","year":"2024","journal-title":"arXiv preprint"},{"key":"ref8","year":"2024","journal-title":"feizhengcong. feizhengcong\/DiT-MoE \u2022 Hugging Face"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00798"},{"key":"ref10","author":"He","year":"2021","journal-title":"FastMoE: A Fast Mixture-of-Expert Training System"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1145\/3503221.3508418"},{"key":"ref12","article-title":"Gans trained by a two time-scale update rule converge to a local nash equilibrium","volume-title":"Advances in Neural Information Processing Systems","author":"Heusel","year":"2017"},{"key":"ref13","author":"Ho","year":"2022","journal-title":"Classifier-free diffusion guidance"},{"key":"ref14","first-page":"6840","article-title":"Denoising diffusion probabilistic models","volume":"33","author":"Ho","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref15","article-title":"Imagen video: High definition video generation with diffusion models","author":"Ho","year":"2022","journal-title":"arXiv preprint"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1714"},{"key":"ref17","article-title":"Improved precision and recall metric for assessing generative models","volume-title":"Advances in Neural Information Processing Systems","author":"Kynk\u00e4\u00e4nniemi","year":"2019"},{"key":"ref18","author":"Lepikhin","year":"2020","journal-title":"GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding"},{"key":"ref19","first-page":"6265","article-title":"Base layers: Simplifying training of large, sparse models","volume-title":"Proceedings of the 38th International Conference on Machine Learning","author":"Lewis","year":"2021"},{"key":"ref20","first-page":"945","article-title":"Accelerating distributed MoE training and inference with lina","volume-title":"2023 USENIX Annual Technical Conference (USENIX ATC 23)","author":"Li"},{"key":"ref21","author":"Li","year":"2024","journal-title":"Locmoe: A low-overhead moe for large language model training"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00686"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.134"},{"key":"ref24","author":"Liu","year":"2022","journal-title":"Flow straight and fast: Learning to generate and transfer data with rectified flow"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM53939.2023.10229053"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2024.3504284"},{"key":"ref27","article-title":"Videofusion: Decomposed diffusion models for high-quality video generation","author":"Luo","year":"2023","journal-title":"arXiv preprint"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.52202\/079017-4236"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01492"},{"key":"ref30","author":"Nash","year":"2021","journal-title":"Generating images with sparse representations"},{"key":"ref31","first-page":"8024","article-title":"Pytorch: An imperative style, high-performance deep learning library","volume-title":"Advances in Neural Information Processing Systems 32","author":"Paszke","year":"2019"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref33","first-page":"1833218346","article-title":"DeepSpeed-MoE: Advancing mixture-of-experts inference and training to power nextgeneration AI scale","volume-title":"Proceedings of the 39th International Conference on Machine Learning","author":"Rajbhandari","year":"2022"},{"key":"ref34","article-title":"Improved techniques for training gans","volume-title":"Advances in Neural Information Processing Systems","author":"Salimans","year":"2016"},{"key":"ref35","author":"Selvaraju","year":"2024","journal-title":"FORA: Fast-Forward Caching in Diffusion Transformer Acceleration"},{"key":"ref36","author":"Shazeer","year":"2017","journal-title":"Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer"},{"key":"ref37","author":"Shoeybi","year":"2020","journal-title":"MegatronLM: Training Multi-Billion Parameter Language Models Using Model Parallelism"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73464-9_20"},{"key":"ref39","first-page":"11287","article-title":"Score-based generative modeling in latent space","volume":"34","author":"Vahdat","year":"2021","journal-title":"Advances in neural information processing systems"},{"key":"ref40","article-title":"Attention is all you need","volume-title":"Advances in Neural Information Processing Systems","author":"Vaswani","year":"2017"},{"key":"ref41","author":"Wang","year":"2024","journal-title":"PipeFusion: Displaced Patch Pipeline Parallelism for Inference of Diffusion Transformer Models"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00594"},{"key":"ref43","article-title":"Raphael: text-to-image generation via large mixture of diffusion paths","volume-title":"Proceedings of the 37th International Conference on Neural Information Processing Systems","author":"Xue"},{"key":"ref44","first-page":"915925","article-title":"Exploiting InterLayer Expert Affinity for Accelerating Mixture-of-Experts Model Inference","volume-title":"2024 IEEE International Parallel and Distributed Processing Symposium (IPDPS)","author":"Yao"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11444122.pdf?arnumber=11444122","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:10:00Z","timestamp":1777612200000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11444122\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":44,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01436","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}