[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-a-new-benchmark-tests-whether-robots-remember-what-they-see":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},7359,"a-new-benchmark-tests-whether-robots-remember-what-they-see","A New Benchmark Tests Whether Robots Remember What They See","A new benchmark called ActiveArena shows robots ace familiar scenes but stumble on new ones, pinpointing memory as the bottleneck in active perception.","A new benchmark says most robots are lousy at remembering what they just looked at.\n\nResearchers built ActiveArena-Sim, a simulator with movable cameras and large workspaces, then used it to create ActiveArena-Bench: 35 tasks across 5 categories that force a robot to look around, probe objects, and hold onto that information across multiple steps rather than react to a single snapshot. Alongside it, the team released ActiveArena-VLA, 13 vision-language-action model configurations built to isolate specific variables - how memory gets written, how much of it a robot can hold, whether proprioception (sensing its own joint positions) helps, and whether a planner should manage memory directly. Every task ships with memory annotations and standardized training data, plus separate in-distribution and out-of-distribution test scenes featuring new backgrounds and distractor objects.\n\nThe headline result is a wide gap between how models perform on scenes they trained on and scenes they haven't seen - a familiar failure mode in robotics research, but one this benchmark is built specifically to measure and diagnose. Bigger memory buffers and proprioceptive input help close that gap, but the biggest gain came from a planner that decides what is worth remembering, which matched top performance using far less memory. That is a useful data point for anyone building real robots: the fix for forgetful machines may not be more storage, but better judgment about what to store.\n\nThis is still lab-only benchmarking, not a robot arm proving it can find your keys in a messy drawer - but it is a sharper test than the static, single-glance setups most manipulation research still leans on.","[\"robotics\",\"ai-research\",\"benchmarks\",\"computer-vision\"]","2026-09-23T04:00:00.000Z","2026-09-23T09:27:36.486Z","2026-09-23T09:27:42.754Z","published",null,[],"ai",[26,27,28,29],"robotics","ai-research","benchmarks","computer-vision",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.24124",0,{"sections":36},[37,40,44,49,54,58,62,67,72,77,82,87,92,97],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",4297,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",710,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",369,"2026-09-23T02:13:52.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",202,"2026-09-22T23:00:04.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":18},"Hardware","hardware",169,{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",133,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",110,"2026-09-22T20:00:00.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Software","software",80,"2026-09-22T23:32:52.000Z",{"name":73,"slug":74,"count":75,"latest_published_at":76},"Dev Tools","dev-tools",79,"2026-09-22T22:21:13.000Z",{"name":78,"slug":79,"count":80,"latest_published_at":81},"Startups","startups",65,"2026-09-22T22:06:48.000Z",{"name":83,"slug":84,"count":85,"latest_published_at":86},"Gaming","gaming",45,"2026-09-22T15:35:06.000Z",{"name":88,"slug":89,"count":90,"latest_published_at":91},"General","general",43,"2026-09-21T23:48:56.000Z",{"name":93,"slug":94,"count":95,"latest_published_at":96},"Reviews","reviews",27,"2026-09-22T13:00:00.000Z",{"name":98,"slug":99,"count":100,"latest_published_at":101},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]