Deprecated: The each() function is deprecated. This message will be suppressed on further calls in /home/zhenxiangba/zhenxiangba.com/public_html/phproxy-improved-master/index.php on line 456 Paper page - GigaBrain-0.5M*: a VLA That Learns From World Model-Based Reinforcement Learning
Please give a thumbs up to this comment if you found it helpful!
\n
If you want recommendations for any Paper on Hugging Face checkout this Space
\n
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend
\n","updatedAt":"2026-02-14T01:40:45.941Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6700775623321533},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2602.12099","authors":[{"_id":"698e8ff2cace060ff123ac59","name":"GigaBrain Team","hidden":false},{"_id":"698e8ff2cace060ff123ac5a","name":"Boyuan Wang","hidden":false},{"_id":"698e8ff2cace060ff123ac5b","name":"Chaojun Ni","hidden":false},{"_id":"698e8ff2cace060ff123ac5c","name":"Guan Huang","hidden":false},{"_id":"698e8ff2cace060ff123ac5d","name":"Guosheng Zhao","hidden":false},{"_id":"698e8ff2cace060ff123ac5e","name":"Hao Li","hidden":false},{"_id":"698e8ff2cace060ff123ac5f","name":"Jie Li","hidden":false},{"_id":"698e8ff2cace060ff123ac60","name":"Jindi Lv","hidden":false},{"_id":"698e8ff2cace060ff123ac61","name":"Jingyu Liu","hidden":false},{"_id":"698e8ff2cace060ff123ac62","name":"Lv Feng","hidden":false},{"_id":"698e8ff2cace060ff123ac63","name":"Mingming Yu","hidden":false},{"_id":"698e8ff2cace060ff123ac64","name":"Peng Li","hidden":false},{"_id":"698e8ff2cace060ff123ac65","name":"Qiuping Deng","hidden":false},{"_id":"698e8ff2cace060ff123ac66","name":"Tianze Liu","hidden":false},{"_id":"698e8ff2cace060ff123ac67","name":"Xinyu Zhou","hidden":false},{"_id":"698e8ff2cace060ff123ac68","name":"Xinze Chen","hidden":false},{"_id":"698e8ff2cace060ff123ac69","user":{"_id":"6426616ea5ec4a5cbc535634","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6426616ea5ec4a5cbc535634/5IfSFYd9QOxz8K9QmBCst.png","isPro":false,"fullname":"JeffWang","user":"Jeff-Wang","type":"user"},"name":"Xiaofeng Wang","status":"claimed_verified","statusLastChangedAt":"2026-02-13T09:36:41.892Z","hidden":false},{"_id":"698e8ff2cace060ff123ac6a","user":{"_id":"644012cf3e0374802e174f7c","avatarUrl":"/avatars/0f4a4bd6f96ce193871843e1d01439e8.svg","isPro":false,"fullname":"Yang Wang","user":"supermodelteam","type":"user"},"name":"Yang Wang","status":"claimed_verified","statusLastChangedAt":"2026-02-13T09:36:39.496Z","hidden":false},{"_id":"698e8ff2cace060ff123ac6b","name":"Yifan Li","hidden":false},{"_id":"698e8ff2cace060ff123ac6c","name":"Yifei Nie","hidden":false},{"_id":"698e8ff2cace060ff123ac6d","name":"Yilong Li","hidden":false},{"_id":"698e8ff2cace060ff123ac6e","name":"Yukun Zhou","hidden":false},{"_id":"698e8ff2cace060ff123ac6f","name":"Yun Ye","hidden":false},{"_id":"698e8ff2cace060ff123ac70","name":"Zhichao Liu","hidden":false},{"_id":"698e8ff2cace060ff123ac71","name":"Zheng Zhu","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/6426616ea5ec4a5cbc535634/w6OBVRe04BfXtAnLDzr3o.mp4"],"publishedAt":"2026-02-12T15:55:19.000Z","submittedOnDailyAt":"2026-02-13T00:24:14.150Z","title":"GigaBrain-0.5M*: a VLA That Learns From World Model-Based Reinforcement Learning","submittedOnDailyBy":{"_id":"6426616ea5ec4a5cbc535634","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6426616ea5ec4a5cbc535634/5IfSFYd9QOxz8K9QmBCst.png","isPro":false,"fullname":"JeffWang","user":"Jeff-Wang","type":"user"},"summary":"Vision-language-action (VLA) models that directly predict multi-step action chunks from current observations face inherent limitations due to constrained scene understanding and weak future anticipation capabilities. In contrast, video world models pre-trained on web-scale video corpora exhibit robust spatiotemporal reasoning and accurate future prediction, making them a natural foundation for enhancing VLA learning. Therefore, we propose GigaBrain-0.5M*, a VLA model trained via world model-based reinforcement learning. Built upon GigaBrain-0.5, which is pre-trained on over 10,000 hours of robotic manipulation data, whose intermediate version currently ranks first on the international RoboChallenge benchmark. GigaBrain-0.5M* further integrates world model-based reinforcement learning via RAMP (Reinforcement leArning via world Model-conditioned Policy) to enable robust cross-task adaptation. Empirical results demonstrate that RAMP achieves substantial performance gains over the RECAP baseline, yielding improvements of approximately 30\\% on challenging tasks including Laundry Folding, Box Packing, and Espresso Preparation. Critically, GigaBrain-0.5M^* exhibits reliable long-horizon execution, consistently accomplishing complex manipulation tasks without failure as validated by real-world deployment videos on our https://gigabrain05m.github.io{project page}.","upvotes":55,"discussionId":"698e8ff2cace060ff123ac72","projectPage":"https://gigabrain05m.github.io/","githubRepo":"https://github.com/open-gigaai/giga-brain-0","githubRepoAddedBy":"user","ai_summary":"A vision-language-action model enhanced with world model-based reinforcement learning demonstrates improved performance and long-horizon execution capabilities for robotic manipulation tasks.","ai_keywords":["Vision-language-action models","world models","reinforcement learning","cross-task adaptation","RAMP","RoboChallenge benchmark","robotic manipulation"],"githubStars":2328,"organization":{"_id":"68d6587936e2de9610d9f5f0","name":"open-gigaai","fullname":"GigaAI","avatar":"https://cdn-uploads.huggingface.co/production/uploads/68d6394328e169473e90e4a6/zUK7FKr_8XqrN0aFUgsD-.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6426616ea5ec4a5cbc535634","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6426616ea5ec4a5cbc535634/5IfSFYd9QOxz8K9QmBCst.png","isPro":false,"fullname":"JeffWang","user":"Jeff-Wang","type":"user"},{"_id":"644012cf3e0374802e174f7c","avatarUrl":"/avatars/0f4a4bd6f96ce193871843e1d01439e8.svg","isPro":false,"fullname":"Yang Wang","user":"supermodelteam","type":"user"},{"_id":"67b04be9e9a7dc7b885a3f00","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/nIm5CkiJ6psrtdjDTN17Z.png","isPro":false,"fullname":"yk202","user":"yk202","type":"user"},{"_id":"63648a9fe31159b7ca6fa576","avatarUrl":"/avatars/d8f1d22cfae52a62aba77f0998371b84.svg","isPro":false,"fullname":"cui","user":"justinbel","type":"user"},{"_id":"67039dea223c62ec88856be3","avatarUrl":"/avatars/8e719779c3581b6041c19602f3883558.svg","isPro":false,"fullname":"Chaojun111","user":"Ni1111","type":"user"},{"_id":"654c7f32386fc5525c1ebc12","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/654c7f32386fc5525c1ebc12/RIUS_nkbt9n5iu46lLQkb.jpeg","isPro":false,"fullname":"zhao guosheng","user":"f1yfish","type":"user"},{"_id":"6884a5478ed76e68274a23b2","avatarUrl":"/avatars/4cf607d2cf418aed84a41942cdd921fc.svg","isPro":false,"fullname":"Hao Li","user":"lh152","type":"user"},{"_id":"656e9b562cd7a3e348011d26","avatarUrl":"/avatars/bcca51bdc27c664f8f132420e6ed99fa.svg","isPro":false,"fullname":"Zheng Zhu","user":"ZhengZhu","type":"user"},{"_id":"64756c98bb0e9dd7764afc24","avatarUrl":"/avatars/eadbaf75cc5a8fe246f836b0a58ea45b.svg","isPro":false,"fullname":"jiagangzhu","user":"hugjiagangzhu","type":"user"},{"_id":"6502ab37b7895899cbe17de8","avatarUrl":"/avatars/ff4b5db80ac99a300b8f873480e70442.svg","isPro":false,"fullname":"Kane Wild","user":"wykgun","type":"user"},{"_id":"662a79fc043e73938e2b94d2","avatarUrl":"/avatars/b7ce0109f02087309da3670029b09396.svg","isPro":false,"fullname":"Mao Jiming","user":"michaelmjm","type":"user"},{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0,"organization":{"_id":"68d6587936e2de9610d9f5f0","name":"open-gigaai","fullname":"GigaAI","avatar":"https://cdn-uploads.huggingface.co/production/uploads/68d6394328e169473e90e4a6/zUK7FKr_8XqrN0aFUgsD-.png"}}">
A vision-language-action model enhanced with world model-based reinforcement learning demonstrates improved performance and long-horizon execution capabilities for robotic manipulation tasks.
AI-generated summary
Vision-language-action (VLA) models that directly predict multi-step action chunks from current observations face inherent limitations due to constrained scene understanding and weak future anticipation capabilities. In contrast, video world models pre-trained on web-scale video corpora exhibit robust spatiotemporal reasoning and accurate future prediction, making them a natural foundation for enhancing VLA learning. Therefore, we propose GigaBrain-0.5M*, a VLA model trained via world model-based reinforcement learning. Built upon GigaBrain-0.5, which is pre-trained on over 10,000 hours of robotic manipulation data, whose intermediate version currently ranks first on the international RoboChallenge benchmark. GigaBrain-0.5M* further integrates world model-based reinforcement learning via RAMP (Reinforcement leArning via world Model-conditioned Policy) to enable robust cross-task adaptation. Empirical results demonstrate that RAMP achieves substantial performance gains over the RECAP baseline, yielding improvements of approximately 30\% on challenging tasks including Laundry Folding, Box Packing, and Espresso Preparation. Critically, GigaBrain-0.5M^* exhibits reliable long-horizon execution, consistently accomplishing complex manipulation tasks without failure as validated by real-world deployment videos on our https://gigabrain05m.github.io{project page}.
GigaBrain-0.5M* is a VLA That Learns From World Model-Based Reinforcement Learning. GigaBrain-0.5M* exhibits reliable long-horizon execution, consistently accomplishing complex manipulation tasks without failure.