[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-a-fix-for-robots-that-lose-their-bearings-mid-task":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},10659,"a-fix-for-robots-that-lose-their-bearings-mid-task","A Fix for Robots That Lose Their Bearings Mid-Task","A new method called SALT lets robot AI models correct themselves on the fly when the camera feed suddenly looks different, without retraining.","Researchers have a new trick for keeping robots from freezing up when their cameras suddenly see the world differently.\n\nThe method, called Self-supervised Adaptation from Leftover Trajectories (SALT), targets vision-language-action models that control robots end to end. Instead of needing labeled examples of what a disruption looks like, SALT uses the unexecuted tail end of the robot's previous planned action - the leftover trajectory - as a built-in training signal. Because action plans overlap in time, that leftover gives the model something to correct toward at the exact moment a visual glitch hits, then keeps propagating that fix forward as the robot keeps moving. A lightweight gate, tuned only on normal footage, decides when this correction should kick in at all.\n\nThe appeal is that SALT needs zero disruption annotations, expert demonstrations, or new training data from the target environment - it bootstraps entirely from the policy's own predictions. On the LIBERO-10 benchmark, it lifted average success across five persistent visual corruptions from 43.9% to 53.2% for SmolVLA and from 58.7% to 66.0% for GR00T N1.7, while leaving normal-condition performance intact.\n\nOn a real robot, task progress climbed from 0.49 to 0.61 under combined digital and physical disruptions - a reminder that most robot demos you see are filmed under ideal lighting, not the glare, dust, and clutter of an actual room.","[\"robotics\",\"vision-language-action-models\",\"self-supervised-learning\",\"ai-research\"]","2026-10-07T04:00:00.000Z","2026-10-09T02:06:45.519Z","2026-10-09T02:06:49.701Z","published",null,[],"ai",[26,27,28,29],"robotics","vision-language-action-models","self-supervised-learning","ai-research",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.07946",0,{"sections":36},[37,41,46,51,56,61,65,70,75,79,84,89,94,99],{"name":38,"slug":24,"count":39,"latest_published_at":40},"AI",6506,"2026-10-07T18:45:00.000Z",{"name":42,"slug":43,"count":44,"latest_published_at":45},"Security","security",911,"2026-10-07T19:53:42.000Z",{"name":47,"slug":48,"count":49,"latest_published_at":50},"Policy","policy",474,"2026-10-07T18:23:21.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":55},"Deals","deals",453,"2026-10-07T23:58:31.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Hardware","hardware",222,"2026-10-07T21:19:54.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":18},"Science","science",187,{"name":66,"slug":67,"count":68,"latest_published_at":69},"Consumer Tech","consumer-tech",174,"2026-10-07T17:41:41.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Software","software",113,"2026-10-07T18:10:00.000Z",{"name":76,"slug":77,"count":73,"latest_published_at":78},"Startups","startups","2026-10-07T23:36:57.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Dev Tools","dev-tools",105,"2026-10-07T16:59:11.000Z",{"name":85,"slug":86,"count":87,"latest_published_at":88},"General","general",61,"2026-10-07T22:00:24.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"Gaming","gaming",56,"2026-10-07T12:00:00.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",33,"2026-10-05T11:57:17.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]