{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T10:09:18Z","timestamp":1777889358104,"version":"3.51.4"},"reference-count":64,"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\/iccv51701.2025.01752","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"18855-18865","source":"Crossref","is-referenced-by-count":0,"title":["GTR: Guided Thought Reinforcement Prevents Thought Collapse in RL-Based VLM Agent Training"],"prefix":"10.1109","author":[{"given":"Tong","family":"Wei","sequence":"first","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yijun","family":"Yang","sequence":"additional","affiliation":[{"name":"Tencent"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Junliang","family":"Xing","sequence":"additional","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuanchun","family":"Shi","sequence":"additional","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zongqing","family":"Lu","sequence":"additional","affiliation":[{"name":"Peking University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Deheng","family":"Ye","sequence":"additional","affiliation":[{"name":"Tencent"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Gpt-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint"},{"key":"ref2","article-title":"Do as i can, not as i say: Grounding language in robotic affordances","author":"Ahn","year":"2022","journal-title":"arXiv preprint"},{"key":"ref3","article-title":"Rt-2: Vision-language-action models transfer web knowledge to robotic control","author":"Brohan","year":"2023","journal-title":"arXiv preprint"},{"key":"ref4","first-page":"3676","article-title":"Grounding large language models in interactive environments with online reinforcement learning","volume-title":"International conference on machine learning","author":"Carta","year":"2023"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0870"},{"key":"ref6","article-title":"Process reinforcement through implicit rewards","author":"Cui","year":"2025","journal-title":"arXiv preprint"},{"key":"ref7","article-title":"Palm-e: An embodied multimodal language model","volume-title":"International conference on machine learning","author":"Driess","year":"2023"},{"key":"ref8","article-title":"Alphazero-like tree-search can guide large language model decoding and training","author":"Feng","year":"2023","journal-title":"arXiv preprint"},{"key":"ref9","first-page":"20132","article-title":"A minimalist approach to offline reinforcement learning","volume":"34","author":"Fujimoto","year":"2021","journal-title":"Advances in neural information processing systems"},{"key":"ref10","article-title":"Llm critics help catch bugs in mathematics: Towards a better mathematical verifier with natural language feedback","author":"Gao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/ICRA57147.2024.10610090"},{"key":"ref12","article-title":"Deepseek-r1: Incentivizing reasoning capability in 11 ms via reinforcement learning","author":"Guo","year":"2025","journal-title":"arXiv preprint"},{"key":"ref13","article-title":"Advancing language model reasoning through reinforcement learning and inference scaling","author":"Hou","year":"2025","journal-title":"arXiv preprint"},{"issue":"2","key":"ref14","first-page":"3","article-title":"Lora: Low-rank adaptation of large language models","volume":"1","author":"Hu","year":"2022","journal-title":"ICLR"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1924"},{"key":"ref16","first-page":"9118","article-title":"Language models as zero-shot planners: Extracting actionable knowledge for embodied agents","volume-title":"International conference on machine learning","author":"Huang","year":"2022"},{"key":"ref17","article-title":"Voxposer: Composable 3d value maps for robotic manipulation with language models","author":"Huang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref18","first-page":"1","article-title":"Guided policy search","volume-title":"International conference on machine learning","author":"Levine","year":"2013"},{"key":"ref19","article-title":"Muep: a multimodal benchmark for embodied planning with foundation models","author":"Li","year":"2024","journal-title":"IJCAI"},{"key":"ref20","article-title":"Let\u2019s verify step by step","volume-title":"The twelfth international conference on learning representations","author":"Lightman","year":"2023"},{"key":"ref21","author":"Liu","year":"2023","journal-title":"Improved baselines with visual instruction tuning"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1516"},{"key":"ref23","author":"Liu","year":"2024","journal-title":"Llava-next: Improved reasoning, ocr, and world knowledge"},{"key":"ref24","article-title":"Improve mathematical reasoning in language models by automated process supervision","volume":"2","author":"Luo","year":"2024","journal-title":"arXiv preprint"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1090"},{"key":"ref26","year":"2024","journal-title":"Hello gpt-4o"},{"key":"ref27","year":"2024","journal-title":"Learning to reason with 11 ms"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2011"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1145\/3586183.3606763"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2338"},{"key":"ref31","article-title":"Is reinforcement learning (not) for natural language processing: Benchmarks, baselines, and building blocks for natural language policy optimization","author":"Ramamurthy","year":"2022","journal-title":"arXiv preprint"},{"key":"ref32","first-page":"627","article-title":"A reduction of imitation learning and structured prediction to no-regret online learning","volume-title":"Proceedings of the fourteenth international conference on artificial intelligence and statistics","author":"Ross","year":"2011"},{"key":"ref33","article-title":"Proximal policy optimization algorithms","author":"Schulman","year":"2017","journal-title":"arXiv preprint"},{"key":"ref34","article-title":"Rewarding progress: Scaling automated process verifiers for 11 m reasoning","author":"Setlur","year":"2024","journal-title":"arXiv preprint"},{"key":"ref35","article-title":"Deepseekmath: Pushing the limits of mathematical reasoning in open language models","author":"Shao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0377"},{"key":"ref37","article-title":"Alfworld: Aligning text and embodied environments for interactive learning","author":"Shridhar","year":"2020","journal-title":"arXiv preprint"},{"key":"ref38","first-page":"3008","article-title":"Learning to summarize with human feedback","volume":"33","author":"Stiennon","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref39","article-title":"Distilling internet-scale visionlanguage models into embodied agents","author":"Sumers","year":"2023","journal-title":"arXiv preprint"},{"key":"ref40","article-title":"Aligning large multimodal models with factually augmented rlhf","author":"Sun","year":"2023","journal-title":"arXiv preprint"},{"key":"ref41","article-title":"Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context","author":"Team","year":"2024","journal-title":"arXiv preprint"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06747-5"},{"key":"ref43","article-title":"Solving math word problems with process-and outcome-based feedback","author":"Uesato","year":"2022","journal-title":"arXiv preprint"},{"key":"ref44","article-title":"Q*: Improving multi-step reasoning for 11 ms with deliberative planning","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref45","article-title":"Voyager: An open-ended embodied agent with large language models","author":"Wang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref46","article-title":"Math-shepherd: A label-free step-by-step verifier for 11 ms in mathematical reasoning","author":"Wang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref47","article-title":"Describe, explain, plan and select: Interactive planning with large language models enables openworld multi-task agents","author":"Wang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1800"},{"key":"ref49","article-title":"Autogen: Enabling next-gen llm applications via multi-agent conversation","author":"Wu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4222-0"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i26.34987"},{"key":"ref52","article-title":"Deepseek-prover-v1. 5: Harnessing proof assistant feedback for reinforcement learning and monte-carlo tree search","author":"Xin","year":"2024","journal-title":"arXiv preprint"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73232-4_2"},{"key":"ref54","article-title":"Foundation models for decision making: Problems, methods, and opportunities","author":"Yang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.02482"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0517"},{"key":"ref57","article-title":"React: Synergizing reasoning and acting in language models","volume-title":"International conference on learning representations","author":"Yao","year":"2023"},{"key":"ref58","article-title":"Demystifying long chain-of-thought reasoning in llms","author":"Yeo","year":"2025","journal-title":"arXiv preprint"},{"key":"ref59","article-title":"Free process rewards without process labels","author":"Yuan","year":"2024","journal-title":"arXiv preprint"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3522"},{"key":"ref61","article-title":"Generative verifiers: Reward modeling as next-token prediction","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref62","article-title":"Wall-e: World alignment by rule learning improves world model-based 11 m agents","author":"Zhou","year":"2024","journal-title":"arXiv preprint"},{"key":"ref63","article-title":"Archer: Training language model agents via hierarchical multi-turn rl","author":"Zhou","year":"2024","journal-title":"arXiv preprint"},{"key":"ref64","article-title":"Fine-tuning language models from human preferences","author":"Ziegler","year":"2019","journal-title":"arXiv preprint"}],"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\/11445972.pdf?arnumber=11445972","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:09:49Z","timestamp":1777612189000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11445972\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":64,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01752","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}