🚀 TacForcing project page and real-world demos are now available!</p>\n<p>TacForcing enables VLA policies to incorporate execution-time tactile feedback without a separate high-frequency reactive controller. It progressively generates and executes action blocks while refining unfinished actions with fresh tactile feedback.</p>\n<p>🤗 Hugging Face demos: <a href=\"https://huggingface.co/spaces/88runaway/tacforcing\">https://huggingface.co/spaces/88runaway/tacforcing</a><br>🌐 Project page: <a href=\"https://88runaway.github.io/tacforcing/\" rel=\"nofollow\">https://88runaway.github.io/tacforcing/</a><br>📄 Paper: <a href=\"https://arxiv.org/abs/2608.25798\" rel=\"nofollow\">https://arxiv.org/abs/2608.25798</a></p>\n<p>Code is coming soon. Feedback and discussion are welcome!</p>\n","updatedAt":"2026-08-28T09:26:57.814Z","author":{"_id":"67dfaa203f6b3fd65ab95530","avatarUrl":"/avatars/8de806669cdcf41b1606f5c7445e403f.svg","fullname":"Ian","name":"88runaway","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7588097453117371},"editors":["88runaway"],"editorAvatarUrls":["/avatars/8de806669cdcf41b1606f5c7445e403f.svg"],"reactions":[],"isReport":false}},{"id":"6a919b42a7bd263a0098012a","author":{"_id":"64bba541da140e461924dfed","avatarUrl":"/avatars/367993765b0ca3734b2b100db33ed787.svg","fullname":"zhijie deng","name":"zhijie3","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false},"createdAt":"2026-08-28T14:29:22.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"https://88runaway.github.io/tacforcing/","html":"<p><a href=\"https://88runaway.github.io/tacforcing/\" rel=\"nofollow\">https://88runaway.github.io/tacforcing/</a></p>\n","updatedAt":"2026-08-28T14:29:22.225Z","author":{"_id":"64bba541da140e461924dfed","avatarUrl":"/avatars/367993765b0ca3734b2b100db33ed787.svg","fullname":"zhijie deng","name":"zhijie3","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5372741222381592},"editors":["zhijie3"],"editorAvatarUrls":["/avatars/367993765b0ca3734b2b100db33ed787.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.25798","authors":[{"_id":"6a914b43a64059bab69c3767","name":"Jianbo Zhou","hidden":false},{"_id":"6a914b43a64059bab69c3768","name":"Boyuan Zhao","hidden":false},{"_id":"6a914b43a64059bab69c3769","name":"Yuzheng Zhang","hidden":false},{"_id":"6a914b43a64059bab69c376a","name":"Yiyang Chen","hidden":false},{"_id":"6a914b43a64059bab69c376b","name":"Wenxin Chen","hidden":false},{"_id":"6a914b43a64059bab69c376c","name":"Qiuyue Li","hidden":false},{"_id":"6a914b43a64059bab69c376d","name":"Xiangyang Gu","hidden":false},{"_id":"6a914b43a64059bab69c376e","name":"Yuhan Cao","hidden":false},{"_id":"6a914b43a64059bab69c376f","name":"Xiao Xia","hidden":false},{"_id":"6a914b43a64059bab69c3770","name":"Yanzhe Hu","hidden":false},{"_id":"6a914b43a64059bab69c3771","name":"Zhijie Deng","hidden":false}],"publishedAt":"2026-08-26T00:00:00.000Z","submittedOnDailyAt":"2026-08-28T00:00:00.000Z","title":"TacForcing: Streaming Action Generation with Execution-Time Tactile Feedback","submittedOnDailyBy":{"_id":"64bba541da140e461924dfed","avatarUrl":"/avatars/367993765b0ca3734b2b100db33ed787.svg","isPro":true,"fullname":"zhijie deng","user":"zhijie3","type":"user","name":"zhijie3"},"summary":"Contact-rich manipulation requires adapting to contact states that can evolve substantially within an action horizon. However, chunk-based vision-language-action models predict complete action chunks from observations collected before execution, leaving tactile conditioning stale during execution. Existing tactile-reactive approaches typically rely on separate high-frequency controllers, which increase both architectural and training complexity. In this paper, we introduce TacForcing, a streaming action-generation framework that effectively incorporates execution-time tactile feedback. Instead of employing a separate reactive controller, TacForcing replaces the standard action expert with a streaming action expert to generate actions conditioned on the evolving tactile observations acquired during execution. TacForcing also introduces Execution-Aware Tactile Attention (EATA), which restricts tactile conditioning to actions nearing execution, thereby reducing the temporal mismatch between tactile acquisition and action execution. Across six simulated UniVTAC tasks and three real-world contact-rich manipulation tasks, TacForcing achieves average success rates of 65% and 69%, respectively, outperforming strong baselines in both settings.","upvotes":3,"discussionId":"6a914b43a64059bab69c3772","ai_summary":"TacForcing is a streaming action-generation framework that integrates real-time tactile feedback during execution via a streaming action expert and execution-aware tactile attention, improving contact-rich manipulation.","ai_keywords":["vision-language-action models","chunk-based action prediction","tactile conditioning","streaming action expert","Execution-Aware Tactile Attention (EATA)","contact-rich manipulation"],"ai_summary_model":"thinkingmachines/Inkling-Small"},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"67dfaa203f6b3fd65ab95530","avatarUrl":"/avatars/8de806669cdcf41b1606f5c7445e403f.svg","isPro":false,"fullname":"Ian","user":"88runaway","type":"user"},{"_id":"6915db7f0a71fd7d5f9b5b33","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/TIWrne-QAZMACW_sHk_fv.jpeg","isPro":false,"fullname":"Yuxuan Liu","user":"Osc7","type":"user"},{"_id":"64bba541da140e461924dfed","avatarUrl":"/avatars/367993765b0ca3734b2b100db33ed787.svg","isPro":true,"fullname":"zhijie deng","user":"zhijie3","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.25798.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2608.25798 in a model README.md to link it from this page.

Cite arxiv.org/abs/2608.25798 in a dataset README.md to link it from this page.