[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-a-new-gradient-trick-smooths-jerky-robot-policies":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},9085,"a-new-gradient-trick-smooths-jerky-robot-policies","A New Gradient Trick Smooths Jerky Robot Policies","DAMPER smooths jerky reinforcement-learning policies by resolving gradient conflicts, cutting oscillation without sacrificing much reward.","A new gradient-control trick smooths out the jerky movements that plague reinforcement-learning robots, without the usual trial-and-error of tuning extra smoothness penalties.\n\nResearchers built a method called DAMPER that blends the standard actor gradient from algorithms like TD3 and SAC with a temporal-consistency gradient meant to reduce oscillation. Instead of just adding the two together and hoping the balance works out, DAMPER checks whether they conflict: if the smoothness gradient pulls against the actor gradient, that opposing piece gets dropped. What is kept gets rescaled relative to the actor gradient's own size, so the smoothing nudge never overwhelms the signal that is actually driving performance. Across six continuous-control tasks and two algorithms, 12 task-backbone combinations in total, this cut action oscillation every time, and produced the best oscillation score among comparison methods in eight of the twelve.\n\nThat conflict check is the real idea here. Auxiliary smoothness losses are common in robotics and control work, but they are notoriously finicky: set the weight too low and they do nothing, set it too high and they fight the main objective. DAMPER sidesteps that balancing act by design rather than by hand-tuned coefficients, which matters for anyone running actuators or motors where jittery commands mean faster wear and higher energy draw.\n\nIt is not a free upgrade, though. The paper reports task-dependent trade-offs on returns, meaning smoother motion sometimes costs a bit of raw performance, and the evaluation stops at simulated benchmarks rather than physical hardware.","[\"reinforcement-learning\",\"robotics\",\"continuous-control\",\"ai-research\"]","2026-10-01T04:00:00.000Z","2026-10-01T19:09:52.976Z","2026-10-01T19:09:56.118Z","published",null,[],"ai",[26,27,28,29],"reinforcement-learning","robotics","continuous-control","ai-research",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.38903",0,{"sections":36},[37,40,44,49,54,59,63,68,73,77,82,87,92,97],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",5496,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",809,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",429,"2026-10-01T02:26:17.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",298,"2026-09-30T21:00:26.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":58},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":60,"slug":61,"count":62,"latest_published_at":18},"Science","science",162,{"name":64,"slug":65,"count":66,"latest_published_at":67},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":69,"slug":70,"count":71,"latest_published_at":72},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":74,"slug":75,"count":71,"latest_published_at":76},"Software","software","2026-09-30T21:41:11.000Z",{"name":78,"slug":79,"count":80,"latest_published_at":81},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":83,"slug":84,"count":85,"latest_published_at":86},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]