{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T17:03:12Z","timestamp":1790269392475,"version":"4.1.0"},"reference-count":74,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"10","license":[{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"funder":[{"name":"NSFC-RGC Joint Research Scheme","award":["N_HKU769\/25"],"award-info":[{"award-number":["N_HKU769\/25"]}]},{"name":"National Nature Science Foundation of China","award":["62561160156"],"award-info":[{"award-number":["62561160156"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2026,10]]},"DOI":"10.1109\/tpami.2026.3700227","type":"journal-article","created":{"date-parts":[[2026,6,4]],"date-time":"2026-06-04T19:55:26Z","timestamp":1780602926000},"page":"12740-12757","source":"Crossref","is-referenced-by-count":0,"title":["SJD++: Improved Speculative Jacobi Decoding for Training-Free Acceleration of Discrete Auto-Regressive Text-to-Image Generation"],"prefix":"10.1109","volume":"48","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-5077-5326","authenticated-orcid":false,"given":"Yao","family":"Teng","sequence":"first","affiliation":[{"name":"The University of Hong Kong (HKU)"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhihuan","family":"Jiang","sequence":"additional","affiliation":[{"name":"The University of Hong Kong (HKU)"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Han","family":"Shi","sequence":"additional","affiliation":[{"name":"Huawei Noah\u2019s Ark Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9817-7418","authenticated-orcid":false,"given":"Xian","family":"Liu","sequence":"additional","affiliation":[{"name":"xAI","place":["Palo Alto, CA, USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2209-8312","authenticated-orcid":false,"given":"Xuefei","family":"Ning","sequence":"additional","affiliation":[{"name":"Tsinghua University","place":["Beijing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8464-0130","authenticated-orcid":false,"given":"Guohao","family":"Dai","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2931-8958","authenticated-orcid":false,"given":"Yu","family":"Wang","sequence":"additional","affiliation":[{"name":"Tsinghua University","place":["Beijing, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8492-3069","authenticated-orcid":false,"given":"Zhenguo","family":"Li","sequence":"additional","affiliation":[{"name":"Huawei Noah\u2019s Ark Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1831-9952","authenticated-orcid":false,"given":"Xihui","family":"Liu","sequence":"additional","affiliation":[{"name":"The University of Hong Kong (HKU)"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Sparks of artificial general intelligence: Early experiments with GPT-4","author":"Bubeck","year":"2023"},{"key":"ref2","article-title":"Scaling autoregressive models for content-rich text-to-image generation","author":"Yu","year":"2022","journal-title":"Trans. Mach. Learn. Res. (TMLR)"},{"key":"ref3","first-page":"25105","article-title":"VideoPoet: A large language model for zero-shot video generation","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Kondratyuk","year":"2024"},{"key":"ref4","article-title":"Loong: Generating minute-level long videos with autoregressive language models","author":"Wang","year":"2024"},{"key":"ref5","first-page":"19822","article-title":"CogView: Mastering text-to-image generation via transformers","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"34","author":"Ding","year":"2021"},{"key":"ref6","first-page":"8821","article-title":"Zero-shot text-to-image generation","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Ramesh","year":"2021"},{"key":"ref7","article-title":"Improving language understanding by generative pre-training","author":"Radford","year":"2018"},{"key":"ref8","article-title":"LLaMA: Open and efficient foundation language models","author":"Touvron","year":"2023"},{"key":"ref9","article-title":"Chameleon: Mixed-modal early-fusion foundation models","author":"Team","year":"2024"},{"key":"ref10","article-title":"Emu3.5: Native multimodal models are world learners","author":"Cui","year":"2025"},{"key":"ref11","article-title":"LongCat-Next: Lexicalizing modalities as discrete tokens","author":"Team","year":"2026"},{"key":"ref12","first-page":"32211","article-title":"Consistency models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Song","year":"2023"},{"key":"ref13","article-title":"Latent consistency models: Synthesizing high-resolution images with few-step inference","author":"Luo","year":"2023"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00632"},{"key":"ref15","first-page":"25426","article-title":"CLLMs: Consistency large language models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Kou","year":"2024"},{"key":"ref16","doi-asserted-by":"crossref","DOI":"10.1137\/1.9780898719468","volume-title":"Iterative Solution of Nonlinear Equations in Several Variables","author":"Ortega","year":"2000"},{"key":"ref17","first-page":"9791","article-title":"Accelerating feedforward computation via parallel nonlinear equation solving","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Song","year":"2021"},{"key":"ref18","doi-asserted-by":"crossref","DOI":"10.1007\/s11263-026-02795-2","article-title":"Lumina-mGPT: Illuminate flexible photorealistic text-to-image generation with multimodal generative pretraining","volume":"134","author":"Liu","year":"2026","journal-title":"Int. J. Comput. Vis. (IJCV)"},{"key":"ref19","article-title":"Anole: An open, autoregressive, native large multimodal models for interleaved image-text generation","author":"Chern","year":"2024"},{"key":"ref20","article-title":"Autoregressive model beats diffusion: Llama for scalable image generation","author":"Sun","year":"2024"},{"key":"ref21","first-page":"19274","article-title":"Fast inference from transformers via speculative decoding","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Leviathan","year":"2023"},{"key":"ref22","article-title":"Accelerating large language model decoding with speculative sampling","author":"Chen","year":"2023"},{"key":"ref23","article-title":"Janus-Pro: Unified multimodal understanding and generation with data and model scaling","author":"Chen","year":"2025"},{"key":"ref24","article-title":"Simplear: Pushing the frontier of autoregressive visual generation through pretraining, SFT, and RL","author":"Wang","year":"2025"},{"key":"ref25","article-title":"Emu3: Next-token prediction is all you need","year":"2024"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2270"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/tpami.2025.3531907"},{"key":"ref29","article-title":"Accelerating auto-regressive text-to-image generation with training-free speculative Jacobi decoding","volume-title":"Proc. Int. Conf. Learn. Representations (ICLR)","author":"Teng","year":"2025"},{"key":"ref30","first-page":"4797","article-title":"Conditional image generation with PixelCNN decoders","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"29","author":"Oord","year":"2016"},{"key":"ref31","article-title":"PixelCNN++: Improving the PixelCNN with discretized logistic mixture likelihood and other modifications","author":"Salimans","year":"2017"},{"key":"ref32","first-page":"864","article-title":"PixelSNAIL: An improved autoregressive generative model","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Chen","year":"2018"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref35","first-page":"17123","article-title":"Mars: Mixture of auto-regressive models for fine-grained text-to-image synthesis","volume-title":"Proc. AAAI Conf. Artif. Intell.","author":"He","year":"2024"},{"key":"ref36","article-title":"Dalle-2","year":"2023"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"ref38","first-page":"12606","article-title":"Scaling rectified flow transformers for high-resolution image synthesis","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Esser","year":"2024"},{"key":"ref39","article-title":"Progressive distillation for fast sampling of diffusion models","volume-title":"Proc. Int. Conf. Learn. Representations (ICLR)","author":"Salimans","year":"2022"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2668"},{"key":"ref41","article-title":"Consistency trajectory models: Learning probability flow ode trajectory of diffusion","volume-title":"Proc. Int. Conf. Learn. Representations (ICLR)","author":"Kim","year":"2024"},{"key":"ref42","first-page":"323","article-title":"Accelerating image generation with sub-path linear approximation model","volume-title":"Proc. Eur. Conf. Comput. Vis.","author":"Xu","year":"2024"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1505"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0037"},{"key":"ref45","first-page":"285","article-title":"MixDQ: Memory-efficient few-step text-to-image diffusion models with metric-decoupled mixed precision quantization","volume-title":"Proc. Eur. Conf. Comput. Vis.","author":"Zhao","year":"2024"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01492"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01103"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2694"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1797"},{"key":"ref50","article-title":"A survey on efficient inference for large language models","author":"Zhou","year":"2024"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.1027"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.52202\/079017-4443"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2279"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.195"},{"key":"ref55","article-title":"DeepSeek-V2: A strong, economical, and efficient mixture-of-experts language model","year":"2024"},{"key":"ref56","article-title":"PyramidKV: Dynamic KV cache compression based on pyramidal information funneling","author":"Zhang","year":"2024"},{"key":"ref57","article-title":"MoA: Mixture of sparse attention for automatic large language model compression","author":"Fu","year":"2024"},{"key":"ref58","first-page":"28480","article-title":"Evaluating quantized large language models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Li","year":"2024"},{"key":"ref59","first-page":"15706","article-title":"Better & faster large language models via multi-token prediction","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Gloeckle","year":"2024"},{"key":"ref60","first-page":"28935","article-title":"Eagle: Speculative sampling requires rethinking feature uncertainty","volume-title":"Proc. Int. Conf. Mach. Learn. (ICML), PMLR","volume":"235","author":"Li","year":"2024"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1314"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.689"},{"key":"ref63","first-page":"14060","article-title":"Break the sequential dependency of LLM inference using lookahead decoding","volume-title":"Proc. Int. Conf. Mach. Learn. (ICML), PMLR","volume":"235","author":"Fu","year":"2024"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.52202\/079017-4116"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.338"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.52202\/075280-3443"},{"key":"ref67","article-title":"Lantern: Accelerating visual autoregressive models with relaxed speculative decoding","volume-title":"Proc. Int. Conf. Learn. Representations (ICLR)","author":"Jang","year":"2025"},{"key":"ref68","first-page":"22368","article-title":"Zipar: Accelerating autoregressive image generation through spatial locality","volume-title":"Proc. Int. Conf. Mach. Learn. (ICML), PMLR","volume":"267","author":"He","year":"2025"},{"key":"ref69","first-page":"6629","article-title":"GANs trained by a two time-scale update rule converge to a local Nash equilibrium","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"30","author":"Heusel","year":"2017"},{"key":"ref70","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"Proc. Int. Conf. Mach. Learn.","volume":"139","author":"Radford","year":"2021"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/iccv51070.2023.00200"},{"key":"ref72","first-page":"690","article-title":"Deep equilibrium models","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"32","author":"Bai","year":"2019"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00070"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00579"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/34\/11674301\/11551514.pdf?arnumber=11551514","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T16:20:41Z","timestamp":1790266841000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11551514\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10]]},"references-count":74,"journal-issue":{"issue":"10"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2026.3700227","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10]]}}}