[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-geoprop-gives-robot-policies-a-sense-of-where-they-are":10,"sections":40},{"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":30,"tags":31,"sources":35,"feedback":39,"feedback_at":22,"cost_usd":39,"total_tokens":39},4496,"geoprop-gives-robot-policies-a-sense-of-where-they-are","GeoProp Gives Robot Policies a Sense of Where They Are","A lightweight adapter from Alibaba DAMO Academy improves robot manipulation performance by grounding joint-angle data directly in camera images.","A small add-on module called GeoProp shows that the way robots fuse body-position data with camera images has been quietly wrong for years.\n\nResearchers at Alibaba DAMO Academy built GeoProp as a plug-and-play adapter that takes raw proprioceptive data — the joint angles and positions that tell a robot where its limbs are — and maps them onto the 2D image the robot sees. Instead of feeding those numbers in as a separate vector alongside visual features, GeoProp projects the robot's state onto the image plane, samples the visual context at that exact spot, and uses a technique called FiLM modulation to let the body-state signal influence nearby image features. It also samples a short predicted future position from recent movement, giving the policy a visual preview of where the arm is heading. Tested across 67 tasks total, the adapter improved Diffusion Policy by 8.7 percentage points on 63 simulation tasks and improved the pi_0 policy by 4.0 points on the RoboTwin benchmark subset, with a 10.6% average gain in real-world trials across both policy families — while adding only 2-3% to parameter count.\n\nThe result matters because it exposes a gap in the standard recipe for building manipulation policies: treating proprioception as a context-free number list apparently costs more than researchers assumed, since GeoProp beats some vision-only baselines that ignore joint data altogether. The fix is also unusually cheap — a small adapter rather than a new architecture means it can slot into existing pipelines.\n\nRobotics is littered with clever modules that benchmark well and then disappear into lab repositories, so the real test will be whether GeoProp generalizes beyond the tasks and robot morphologies tested here.","[\"robotics\",\"ai\",\"computer-vision\",\"research\"]","2026-07-09T04:00:00.000Z","2026-07-09T05:42:48.672Z","2026-07-09T05:42:51.571Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The body states '63 simulation tasks' in both the Diffusion Policy claim and the total task count, but the source specifies '67 tasks' total with the Diffusion Policy result covering '63 simulation tasks' — the article conflates the two figures and should distinguish them to avoid an internally inconsistent numerical claim.","resolved","ai",[32,30,33,34],"robotics","computer-vision","research",[36],{"name":37,"url":38},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.07101",0,{"sections":41},[42,46,51,56,61,66,71,76,81,86,90,94,99,104],{"name":43,"slug":30,"count":44,"latest_published_at":45},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":47,"slug":48,"count":49,"latest_published_at":50},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":55},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":65},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":18},"Gaming","gaming",41,{"name":91,"slug":92,"count":89,"latest_published_at":93},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":105,"slug":106,"count":107,"latest_published_at":108},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]