BridgeVLA++ achieves state of the art across five manipulation benchmarks with only 9.2% additional parameters, adding powerful spatio-temporal memory while fully preserving—and often improving—BridgeVLA’s exceptional data efficiency and OOD generalization.</p>\n","updatedAt":"2026-08-06T02:40:49.226Z","author":{"_id":"6337e04b171879571956212f","avatarUrl":"/avatars/aa69142bf92e4c50b192463490251ed9.svg","fullname":"Li Peiyan","name":"LPY","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8123034834861755},"editors":["LPY"],"editorAvatarUrls":["/avatars/aa69142bf92e4c50b192463490251ed9.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.05042","authors":[{"_id":"6a73efcbc5e410d076869a6e","name":"Peiyan Li","hidden":false},{"_id":"6a73efcbc5e410d076869a6f","name":"Yuze Zhu","hidden":false},{"_id":"6a73efcbc5e410d076869a70","name":"Yixiang Chen","hidden":false},{"_id":"6a73efcbc5e410d076869a71","name":"Qisen Ma","hidden":false},{"_id":"6a73efcbc5e410d076869a72","name":"Yuan Xu","hidden":false},{"_id":"6a73efcbc5e410d076869a73","name":"Jiabing Yang","hidden":false},{"_id":"6a73efcbc5e410d076869a74","name":"He Guan","hidden":false},{"_id":"6a73efcbc5e410d076869a75","name":"Yan Huang","hidden":false},{"_id":"6a73efcbc5e410d076869a76","name":"Hongtao Wu","hidden":false},{"_id":"6a73efcbc5e410d076869a77","name":"Xiao Ma","hidden":false},{"_id":"6a73efcbc5e410d076869a78","name":"Tao Kong","hidden":false},{"_id":"6a73efcbc5e410d076869a79","name":"Liang Wang","hidden":false},{"_id":"6a73efcbc5e410d076869a7a","name":"Tieniu Tan","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/6337e04b171879571956212f/ZL4rSyO6E5ztWN0cvMssN.mp4"],"publishedAt":"2026-08-05T00:00:00.000Z","submittedOnDailyAt":"2026-08-06T00:00:00.000Z","title":"BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation","submittedOnDailyBy":{"_id":"6337e04b171879571956212f","avatarUrl":"/avatars/aa69142bf92e4c50b192463490251ed9.svg","isPro":false,"fullname":"Li Peiyan","user":"LPY","type":"user","name":"LPY"},"summary":"Leveraging pre-trained vision-language models (VLMs) to construct vision-language-action (VLA) models has emerged as a promising paradigm for 3D robot manipulation. However, existing 3D VLA methods remain data-hungry, exhibit limited generalization under distribution shifts, and lack explicit memory of past observations. These limitations hinder their application to data-scarce, open-world, and memory-dependent manipulation scenarios. Our previous work, BridgeVLA, improves data efficiency and generalization by preserving the input--output alignment of a pre-trained VLM during 3D action learning: raw point clouds are projected into multi-view images, and intermediate heatmaps are predicted before generating robot actions. In this work, we develop BridgeVLA++ by equipping BridgeVLA with a unified spatio-temporal memory architecture that models persistent spatial context and temporal interaction history. The resulting memory-augmented framework can reason over observation histories while preserving BridgeVLA's data efficiency and generalization capabilities. Extensive experiments show that our framework achieves strong performance on spatial manipulation tasks while exhibiting robust generalization. BridgeVLA++ further achieves state-of-the-art performance on two challenging memory-dependent manipulation benchmarks without sacrificing the data efficiency and generalization of the original BridgeVLA. In addition, BridgeVLA++ performs effectively in bimanual manipulation settings and is validated on an additional real-world robotic platform, demonstrating its scalability across tasks, environments, and robotic platforms. These results establish BridgeVLA++ as a unified 3D vision-language-action framework that simultaneously supports data-efficient learning, robust generalization, and effective memory-aware robot manipulation. Project website: https://bridgevla-plus.github.io/.","upvotes":6,"discussionId":"6a73efcbc5e410d076869a7b","projectPage":"https://bridgevla-plus.github.io/"},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6337e04b171879571956212f","avatarUrl":"/avatars/aa69142bf92e4c50b192463490251ed9.svg","isPro":false,"fullname":"Li Peiyan","user":"LPY","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"6a6aa11e9859d6ac81f379dd","avatarUrl":"/avatars/677da06f7762a169543cf98dfecfdbd6.svg","isPro":false,"fullname":"Andrew Rodriguez","user":"Velvet-Andrew","type":"user"},{"_id":"6a6c7f62ef16c968823e3b4d","avatarUrl":"/avatars/42a1f19bd0ad542978225ebdb71b944d.svg","isPro":false,"fullname":"Brian Wilson","user":"Brian-Wilson","type":"user"},{"_id":"6a6aa0c5b58832f7d0fd5238","avatarUrl":"/avatars/419fbadce90b1e9ad4328b13fb77e8a8.svg","isPro":false,"fullname":"Patricia Harris","user":"Patricia-Harris","type":"user"},{"_id":"6a701d76e8095b6796d346d7","avatarUrl":"/avatars/aec06bc897c4b98b3aef783d8183815b.svg","isPro":false,"fullname":"Joshua White","user":"david-4595153","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.05042.md","query":{}}">
BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation
Abstract
Leveraging pre-trained vision-language models (VLMs) to construct vision-language-action (VLA) models has emerged as a promising paradigm for 3D robot manipulation. However, existing 3D VLA methods remain data-hungry, exhibit limited generalization under distribution shifts, and lack explicit memory of past observations. These limitations hinder their application to data-scarce, open-world, and memory-dependent manipulation scenarios. Our previous work, BridgeVLA, improves data efficiency and generalization by preserving the input--output alignment of a pre-trained VLM during 3D action learning: raw point clouds are projected into multi-view images, and intermediate heatmaps are predicted before generating robot actions. In this work, we develop BridgeVLA++ by equipping BridgeVLA with a unified spatio-temporal memory architecture that models persistent spatial context and temporal interaction history. The resulting memory-augmented framework can reason over observation histories while preserving BridgeVLA's data efficiency and generalization capabilities. Extensive experiments show that our framework achieves strong performance on spatial manipulation tasks while exhibiting robust generalization. BridgeVLA++ further achieves state-of-the-art performance on two challenging memory-dependent manipulation benchmarks without sacrificing the data efficiency and generalization of the original BridgeVLA. In addition, BridgeVLA++ performs effectively in bimanual manipulation settings and is validated on an additional real-world robotic platform, demonstrating its scalability across tasks, environments, and robotic platforms. These results establish BridgeVLA++ as a unified 3D vision-language-action framework that simultaneously supports data-efficient learning, robust generalization, and effective memory-aware robot manipulation. Project website: https://bridgevla-plus.github.io/.
Community
BridgeVLA++ achieves state of the art across five manipulation benchmarks with only 9.2% additional parameters, adding powerful spatio-temporal memory while fully preserving—and often improving—BridgeVLA’s exceptional data efficiency and OOD generalization.
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/2608.05042 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.05042 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.05042 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.