We present PhysCaP, a Physics-Informed Code-as-Policy agent for active perception in robotic manipulation. While vision-language-action policies excel at imitating demonstrations, they rely on passive observation and fail to infer latent physical properties critical for manipulation. PhysCaP augments code-as-policy frameworks with a physics-informed exploration layer that enables explicit information-seeking through interaction. It introduces training-free physical property extraction modules that estimate object mass and stiffness from robot proprioception without additional sensors. To balance exploration costs and the efficiency of information obtained, PhysCaP employs a dual-agent design: a Planner that decides when to explore and when to stop, and a Prioritizer that filters implausible interactions and ranks the remainder using a heuristic priority score, enabling efficient, targeted exploration.</p>\n","updatedAt":"2026-08-24T16:44:04.442Z","author":{"_id":"64ae22dd1aee69ece065cdcd","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64ae22dd1aee69ece065cdcd/JG7QaHIrr4i2k4uwR4pZK.png","fullname":"Min-Hung Chen","name":"cmhungsteve","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":21,"isUserFollowing":false,"primaryOrg":{"avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df9200dc3292a8983e5017/Vs5FPVCH-VZBipV3qKTuy.png","fullname":"NVIDIA","name":"nvidia","type":"org","isHf":false,"plan":"plus"}}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7771667838096619},"editors":["cmhungsteve"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/64ae22dd1aee69ece065cdcd/JG7QaHIrr4i2k4uwR4pZK.png"],"reactions":[],"isReport":false}},{"id":"6a8cf0a530fff78bebee8a46","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":378,"isUserFollowing":false},"createdAt":"2026-08-25T01:32:21.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agents](https://huggingface.co/papers/2607.08448) (2026)\n* [ACE: Agentic Control for Embodied Manipulation via Zero-shot Workflow Reasoning](https://huggingface.co/papers/2607.04162) (2026)\n* [RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents](https://huggingface.co/papers/2607.27881) (2026)\n* [A Few Words Go a Long Way: Language Guided Robot Policy Synthesis](https://huggingface.co/papers/2607.23784) (2026)\n* [IMBench: A Benchmark for Intuitive Robotic Manipulation](https://huggingface.co/papers/2607.15641) (2026)\n* [RoboTALES: Learning Reasoning-Guided Robot Policies via Task-Aligned Simulated Futures](https://huggingface.co/papers/2607.06018) (2026)\n* [World Action Planner: Generalizable Decision-Making with Action-Conditioned World Models](https://huggingface.co/papers/2607.27599) (2026)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"<p>This is an automated message from the <a href=\"https://huggingface.co/librarian-bots\">Librarian Bot</a>. I found the following papers similar to this paper. </p>\n<p>The following papers were recommended by the Semantic Scholar API </p>\n<ul>\n<li><a href=\"https://huggingface.co/papers/2607.08448\">Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agents</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.04162\">ACE: Agentic Control for Embodied Manipulation via Zero-shot Workflow Reasoning</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.27881\">RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.23784\">A Few Words Go a Long Way: Language Guided Robot Policy Synthesis</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.15641\">IMBench: A Benchmark for Intuitive Robotic Manipulation</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.06018\">RoboTALES: Learning Reasoning-Guided Robot Policies via Task-Aligned Simulated Futures</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.27599\">World Action Planner: Generalizable Decision-Making with Action-Conditioned World Models</a> (2026)</li>\n</ul>\n<p> Please give a thumbs up to this comment if you found it helpful!</p>\n<p> If you want recommendations for any Paper on Hugging Face checkout <a href=\"https://huggingface.co/spaces/librarian-bots/recommend_similar_papers\">this</a> Space</p>\n<p> You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: <code>@librarian-bot recommend</code></p>\n","updatedAt":"2026-08-25T01:32:21.252Z","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":378,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7236434817314148},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.21031","authors":[{"_id":"6a8c747f8dd056518b7f5301","name":"Chen-Yu Lin","hidden":false},{"_id":"6a8c747f8dd056518b7f5302","name":"Jing-Wen Chen","hidden":false},{"_id":"6a8c747f8dd056518b7f5303","name":"Hsueh-En Chang","hidden":false},{"_id":"6a8c747f8dd056518b7f5304","name":"Hung-An Chen","hidden":false},{"_id":"6a8c747f8dd056518b7f5305","name":"Sheng-Hsun Chang","hidden":false},{"_id":"6a8c747f8dd056518b7f5306","name":"Chi-Pin Huang","hidden":false},{"_id":"6a8c747f8dd056518b7f5307","name":"Fu-En Yang","hidden":false},{"_id":"6a8c747f8dd056518b7f5308","user":{"_id":"64ae22dd1aee69ece065cdcd","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64ae22dd1aee69ece065cdcd/JG7QaHIrr4i2k4uwR4pZK.png","isPro":false,"fullname":"Min-Hung Chen","user":"cmhungsteve","type":"user","name":"cmhungsteve"},"name":"Min-Hung Chen","status":"claimed_verified","statusLastChangedAt":"2026-08-24T16:45:46.906Z","hidden":false},{"_id":"6a8c747f8dd056518b7f5309","name":"Yi-Ting Chen","hidden":false},{"_id":"6a8c747f8dd056518b7f530a","name":"Yu-Chiang Frank Wang","hidden":false},{"_id":"6a8c747f8dd056518b7f530b","name":"Shao-Hua Sun","hidden":false}],"publishedAt":"2026-08-21T00:00:00.000Z","submittedOnDailyAt":"2026-08-24T00:00:00.000Z","title":"PhysCaP: Grounding Code-as-Policy Agent with Physics-Informed Exploration","submittedOnDailyBy":{"_id":"64ae22dd1aee69ece065cdcd","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64ae22dd1aee69ece065cdcd/JG7QaHIrr4i2k4uwR4pZK.png","isPro":false,"fullname":"Min-Hung Chen","user":"cmhungsteve","type":"user","name":"cmhungsteve"},"summary":"We present PhysCaP, a Physics-Informed Code-as-Policy agent for active perception in robotic manipulation. While vision-language-action policies excel at imitating demonstrations, they rely on passive observation and fail to infer latent physical properties critical for manipulation. PhysCaP augments code-as-policy frameworks with a physics-informed exploration layer that enables explicit information-seeking through interaction. It introduces training-free physical property extraction modules that estimate object mass and stiffness from robot proprioception without additional sensors. To balance exploration costs and the efficiency of information obtained, PhysCaP employs a dual-agent design: a Planner that decides when to explore and when to stop, and a Prioritizer that filters implausible interactions and ranks the remainder using a heuristic priority score, enabling efficient, targeted exploration. We evaluate PhysCaP on real-world tabletop manipulation tasks (searching for hidden objects, detecting empty cans, and finding ripe avocados) and a simulated task in LIBERO. The results show that existing passive and naive interactive baselines either fail when physical properties are hidden or over-explore, whereas PhysCaP achieves comparable performance with fewer interactions and reduced execution time. Ablation studies further validate the effectiveness of the proposed physical property extraction modules. Project page: https://physcap.github.io","upvotes":3,"discussionId":"6a8c747f8dd056518b7f530c","projectPage":"https://physcap.github.io/","ai_summary":"PhysCaP is a physics-informed code-generation agent that actively explores objects to infer hidden physical properties for efficient robotic manipulation.","ai_keywords":["Physics-Informed Code-as-Policy","vision-language-action policies","code-as-policy","physical property extraction","proprioception","dual-agent design","Planner","Prioritizer","LIBERO"],"ai_summary_model":"thinkingmachines/Inkling-Small"},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"64ae22dd1aee69ece065cdcd","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64ae22dd1aee69ece065cdcd/JG7QaHIrr4i2k4uwR4pZK.png","isPro":false,"fullname":"Min-Hung Chen","user":"cmhungsteve","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":"631e14ac473a6825f285e89d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/631e14ac473a6825f285e89d/K-6QnoeGLg8XFvbTMMdqA.jpeg","isPro":false,"fullname":"Yury Panikov","user":"panikov","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.21031.md","query":{}}">
PhysCaP: Grounding Code-as-Policy Agent with Physics-Informed Exploration
Abstract
PhysCaP is a physics-informed code-generation agent that actively explores objects to infer hidden physical properties for efficient robotic manipulation.
We present PhysCaP, a Physics-Informed Code-as-Policy agent for active perception in robotic manipulation. While vision-language-action policies excel at imitating demonstrations, they rely on passive observation and fail to infer latent physical properties critical for manipulation. PhysCaP augments code-as-policy frameworks with a physics-informed exploration layer that enables explicit information-seeking through interaction. It introduces training-free physical property extraction modules that estimate object mass and stiffness from robot proprioception without additional sensors. To balance exploration costs and the efficiency of information obtained, PhysCaP employs a dual-agent design: a Planner that decides when to explore and when to stop, and a Prioritizer that filters implausible interactions and ranks the remainder using a heuristic priority score, enabling efficient, targeted exploration. We evaluate PhysCaP on real-world tabletop manipulation tasks (searching for hidden objects, detecting empty cans, and finding ripe avocados) and a simulated task in LIBERO. The results show that existing passive and naive interactive baselines either fail when physical properties are hidden or over-explore, whereas PhysCaP achieves comparable performance with fewer interactions and reduced execution time. Ablation studies further validate the effectiveness of the proposed physical property extraction modules. Project page: https://physcap.github.io
Community
We present PhysCaP, a Physics-Informed Code-as-Policy agent for active perception in robotic manipulation. While vision-language-action policies excel at imitating demonstrations, they rely on passive observation and fail to infer latent physical properties critical for manipulation. PhysCaP augments code-as-policy frameworks with a physics-informed exploration layer that enables explicit information-seeking through interaction. It introduces training-free physical property extraction modules that estimate object mass and stiffness from robot proprioception without additional sensors. To balance exploration costs and the efficiency of information obtained, PhysCaP employs a dual-agent design: a Planner that decides when to explore and when to stop, and a Prioritizer that filters implausible interactions and ranks the remainder using a heuristic priority score, enabling efficient, targeted exploration.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
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/2608.21031 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.21031 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.21031 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.