{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T10:06:06Z","timestamp":1777889166848,"version":"3.51.4"},"reference-count":34,"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"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01426","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"15375-15384","source":"Crossref","is-referenced-by-count":0,"title":["Grouped Speculative Decoding for Autoregressive Image Generation"],"prefix":"10.1109","author":[{"given":"Junhyuk","family":"So","sequence":"first","affiliation":[{"name":"Graduate School of Artificial Intelligence, POSTECH"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Juncheol","family":"Shin","sequence":"additional","affiliation":[{"name":"Graduate School of Artificial Intelligence, POSTECH"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hyunho","family":"Kook","sequence":"additional","affiliation":[{"name":"Graduate School of Artificial Intelligence, POSTECH"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Eunhyeok","family":"Park","sequence":"additional","affiliation":[{"name":"POSTECH,Department of Computer Science and Engineering"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Gpt-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint arXiv"},{"key":"ref2","article-title":"Cosmos world foundation model platform for physical ai","author":"Agarwal","year":"2025","journal-title":"arXiv preprint arXiv"},{"key":"ref3","article-title":"Qwen technical report","author":"Bai","year":"2023","journal-title":"arXiv preprint arXiv"},{"key":"ref4","article-title":"Dynamic depth decoding: Faster speculative decoding for llms","author":"Brown","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref5","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref6","article-title":"Medusa: Simple 11m inference acceleration framework with multiple decoding heads","author":"Cai","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.02173"},{"key":"ref8","first-page":"6840","article-title":"Denoising diffusion probabilistic models","volume":"33","author":"Ho","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref9","article-title":"Lantern: Accelerating visual autoregressive models with relaxed speculative decoding","author":"Jang","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref10","first-page":"19274","article-title":"Fast inference from transformers via speculative decoding","volume-title":"International Conference on Machine Learning","author":"Leviathan"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref12","article-title":"Lumina-mgpt: Illuminate flexible photorealistic text-to-image generation with multimodal generative pretraining","author":"Liu","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1038\/s41587-022-01618-2"},{"key":"ref14","article-title":"Pytorch: An imperative style, high-performance deep learning library","volume":"32","author":"Paszke","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref16","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"International conference on machine learning","author":"Radford"},{"key":"ref17","article-title":"Denoising diffusion implicit models","author":"Song","year":"2020","journal-title":"arXiv preprint arXiv"},{"key":"ref18","first-page":"9791","article-title":"Accelerating feedforward computation via parallel nonlinear equation solving","volume-title":"International Conference on Machine Learning","author":"Song"},{"key":"ref19","article-title":"Autoregressive model beats diffusion: Llama for scalable image generation","author":"Sun","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1314"},{"key":"ref21","article-title":"Chameleon: Mixed-modal early-fusion foundation models","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref22","article-title":"Accelerating autoregressive text-to-image generation with training-free speculative jacobi decoding","author":"Teng","year":"2024","journal-title":"arXiv preprint arXiv"},{"key":"ref23","article-title":"Llama: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint arXiv"},{"key":"ref24","volume-title":"An optimal lossy variant of speculative decoding","author":"Tran-Thien","year":"2023"},{"key":"ref25","article-title":"Neural discrete representation learning","volume":"30","author":"Van Den Oord","year":"2017","journal-title":"Advances in neural information processing systems"},{"issue":"11","key":"ref26","article-title":"Visualizing data using t -sne","volume":"9","author":"Van der Maaten","year":"2008","journal-title":"Journal of machine learning research"},{"key":"ref27","article-title":"Neural codec language models are zero-shot text to speech synthesizers","author":"Wang","year":"2023","journal-title":"arXiv preprint arXiv"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01209"},{"key":"ref29","article-title":"Continuous speculative decoding for autoregressive image generation","author":"Wang","year":"2024","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.52202\/079017-4067"},{"key":"ref31","article-title":"Vector-quantized image modeling with improved vqgan","author":"Yu","year":"2021","journal-title":"arXiv preprint arXiv"},{"issue":"3","key":"ref32","first-page":"5","article-title":"Scaling autoregressive models for content-rich text-to-image generation","volume":"2","author":"Yu","year":"2022","journal-title":"arXiv preprint arXiv"},{"key":"ref33","article-title":"Speeding up speculative decoding via approximate verification","author":"Zhong","year":"2025","journal-title":"arXiv preprint arXiv"},{"key":"ref34","article-title":"Scaling the codebook size of vqgan to 100,000 with a utilization rate of 99 %","author":"Zhu","year":"2024","journal-title":"arXiv preprint arXiv"}],"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\/11444739.pdf?arnumber=11444739","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:07:24Z","timestamp":1777612044000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11444739\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":34,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01426","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}