We introduce RISE, an adaptive imagination framework for World Action Models (WAMs) in autonomous driving. Instead of using a fixed rollout depth for every scene, RISE dynamically decides whether to continue imagining future latent states based on planning benefit, revealed risk, and computation cost.</p>\n","updatedAt":"2026-08-25T02:22:52.210Z","author":{"_id":"671618d2c6f6570d4f513004","avatarUrl":"/avatars/6ac6577d110731a1e8e3d56b9018fc86.svg","fullname":"LiangYao","name":"1e12Leon","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8119101524353027},"editors":["1e12Leon"],"editorAvatarUrls":["/avatars/6ac6577d110731a1e8e3d56b9018fc86.svg"],"reactions":[],"isReport":false}},{"id":"6a8d09edfe2fe43b7266608e","author":{"_id":"6a0a83ee4b0e1e375a9bb243","avatarUrl":"/avatars/ee8ff7722719bd13879e64539326889a.svg","fullname":"YANG","name":"AlanoneOK","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false},"createdAt":"2026-08-25T03:20:13.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"excellent job","html":"<p>excellent job</p>\n","updatedAt":"2026-08-25T03:20:13.323Z","author":{"_id":"6a0a83ee4b0e1e375a9bb243","avatarUrl":"/avatars/ee8ff7722719bd13879e64539326889a.svg","fullname":"YANG","name":"AlanoneOK","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9151433706283569},"editors":["AlanoneOK"],"editorAvatarUrls":["/avatars/ee8ff7722719bd13879e64539326889a.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.20430","authors":[{"_id":"6a8cedb18dd056518b7f53eb","name":"Hongbo Lu","hidden":false},{"_id":"6a8cedb18dd056518b7f53ec","name":"Liang Yao","hidden":false},{"_id":"6a8cedb18dd056518b7f53ed","name":"Chenghao He","hidden":false},{"_id":"6a8cedb18dd056518b7f53ee","name":"Hao Han","hidden":false},{"_id":"6a8cedb18dd056518b7f53ef","name":"Fan Liu","hidden":false},{"_id":"6a8cedb18dd056518b7f53f0","name":"Wenlong Liao","hidden":false},{"_id":"6a8cedb18dd056518b7f53f1","name":"Tao He","hidden":false},{"_id":"6a8cedb18dd056518b7f53f2","name":"Pai Peng","hidden":false}],"publishedAt":"2026-08-20T00:00:00.000Z","submittedOnDailyAt":"2026-08-25T00:00:00.000Z","title":"RISE: Adaptive Imagination for World Action Models","submittedOnDailyBy":{"_id":"671618d2c6f6570d4f513004","avatarUrl":"/avatars/6ac6577d110731a1e8e3d56b9018fc86.svg","isPro":false,"fullname":"LiangYao","user":"1e12Leon","type":"user","name":"1e12Leon"},"summary":"World Action Models (WAMs) improve planning by incorporating future world evolution into action generation, yet existing methods allocate a fixed imagination budget to every scene. We propose RISE (Refining Imagination through SElective Rollout), a system-level adaptive imagination framework that makes sequential Roll/Stop decisions according to the expected planning benefit of continued rollout. At each step, a Latent Evaluator estimates the risk revealed by the current prefix and how much planning could improve if imagination continues, while a Rollout Gate weighs this expected benefit against additional computation cost. Since factual driving logs expose only one realized future, we further construct CounterDrive, a counterfactual dataset with diverse outcomes and risk levels, to enrich future dynamics and provide localized risk supervision. Each retained sample undergoes expert verification and annotation of trajectory validity, incident onset, and causal category, providing a reusable resource for safety-critical world-modeling research. Experiments on NAVSIM and nuScenes show that RISE achieves the best overall planning performance while reducing unnecessary rollout, with additional transfer results supporting its plug-in generality across WAM architectures.","upvotes":23,"discussionId":"6a8cedb18dd056518b7f53f3","projectPage":"https://cowarobot-ai.github.io/RISE/","githubRepo":"https://github.com/COOWAI/RISE","githubRepoAddedBy":"user","ai_summary":"RISE adaptively decides when to continue or stop imagination rollouts for planning by weighing expected benefit against cost, supported by a counterfactual driving dataset with expert annotations.","ai_keywords":["World Action Models","adaptive imagination","Roll/Stop decisions","Latent Evaluator","Rollout Gate","CounterDrive","counterfactual dataset","risk supervision","NAVSIM","nuScenes"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":20,"organization":{"_id":"6a056ac51ccc9ecd592d8241","name":"COWARobot","fullname":"COWARobot","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/671618d2c6f6570d4f513004/Vc-gBeokccKOOhMP1GBsZ.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"671618d2c6f6570d4f513004","avatarUrl":"/avatars/6ac6577d110731a1e8e3d56b9018fc86.svg","isPro":false,"fullname":"LiangYao","user":"1e12Leon","type":"user"},{"_id":"63ac5701c21e60a3e9b58aa7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ac5701c21e60a3e9b58aa7/g6EX7diOpuA94R2ab-rZC.png","isPro":true,"fullname":"Dipankar Sarkar","user":"dipankarsarkar","type":"user"},{"_id":"67b6f6fa1651277fa0ce8b96","avatarUrl":"/avatars/b41921d4dadeff84356e7e1b25b6099e.svg","isPro":false,"fullname":"Pai Peng","user":"Popeyepeng","type":"user"},{"_id":"67c8191e5999e7df91a4696b","avatarUrl":"/avatars/95efe5b7392b47d5661c75b35db8f053.svg","isPro":false,"fullname":"lll","user":"hahayaayaa","type":"user"},{"_id":"6a0576273fcbc936eefca756","avatarUrl":"/avatars/352b39954ffa3d161d0ec2b6591a9c19.svg","isPro":false,"fullname":"Haoyu Wang","user":"0w0h0y","type":"user"},{"_id":"6a057f5254ba3c3be9003345","avatarUrl":"/avatars/64380875a35886ad12962042a485ba87.svg","isPro":false,"fullname":"CosmoXu","user":"Cosmo1210","type":"user"},{"_id":"6625319be9c66073a4583589","avatarUrl":"/avatars/6452fab90ba44f131f069deee33c1552.svg","isPro":false,"fullname":"kwongfei","user":"coffee666","type":"user"},{"_id":"67f10d6a84486b4ad2f5801f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/mD2B9tGzMi_JT8j4RDP1E.png","isPro":false,"fullname":"Ziyun Chen","user":"saoqiczy","type":"user"},{"_id":"6a1968992e5d6aa2297ee57a","avatarUrl":"/avatars/ad97f33c404ac41c80ee0e10ee570865.svg","isPro":false,"fullname":"wong","user":"prettytomoa","type":"user"},{"_id":"67a4ad31522cb52117e81f08","avatarUrl":"/avatars/643fd04189318f9446d4237a5b0422bf.svg","isPro":false,"fullname":"hanhao","user":"cowa1997","type":"user"},{"_id":"6a0a83ee4b0e1e375a9bb243","avatarUrl":"/avatars/ee8ff7722719bd13879e64539326889a.svg","isPro":false,"fullname":"YANG","user":"AlanoneOK","type":"user"},{"_id":"6a06f8a8c1d7d248d2431aa3","avatarUrl":"/avatars/40efcb7edb163cb364771d4889f5ecad.svg","isPro":false,"fullname":"Yueming Li","user":"yueming2026","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6a056ac51ccc9ecd592d8241","name":"COWARobot","fullname":"COWARobot","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/671618d2c6f6570d4f513004/Vc-gBeokccKOOhMP1GBsZ.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.20430.md","query":{}}">
RISE: Adaptive Imagination for World Action Models
Abstract
RISE adaptively decides when to continue or stop imagination rollouts for planning by weighing expected benefit against cost, supported by a counterfactual driving dataset with expert annotations.
World Action Models (WAMs) improve planning by incorporating future world evolution into action generation, yet existing methods allocate a fixed imagination budget to every scene. We propose RISE (Refining Imagination through SElective Rollout), a system-level adaptive imagination framework that makes sequential Roll/Stop decisions according to the expected planning benefit of continued rollout. At each step, a Latent Evaluator estimates the risk revealed by the current prefix and how much planning could improve if imagination continues, while a Rollout Gate weighs this expected benefit against additional computation cost. Since factual driving logs expose only one realized future, we further construct CounterDrive, a counterfactual dataset with diverse outcomes and risk levels, to enrich future dynamics and provide localized risk supervision. Each retained sample undergoes expert verification and annotation of trajectory validity, incident onset, and causal category, providing a reusable resource for safety-critical world-modeling research. Experiments on NAVSIM and nuScenes show that RISE achieves the best overall planning performance while reducing unnecessary rollout, with additional transfer results supporting its plug-in generality across WAM architectures.
Community
We introduce RISE, an adaptive imagination framework for World Action Models (WAMs) in autonomous driving. Instead of using a fixed rollout depth for every scene, RISE dynamically decides whether to continue imagining future latent states based on planning benefit, revealed risk, and computation cost.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.20430 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.20430 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.20430 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.