[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-toffee-trains-data-agents-on-synthetic-analytical-paths":10,"sections":34},{"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":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},4292,"toffee-trains-data-agents-on-synthetic-analytical-paths","TOFFEE Trains Data Agents on Synthetic Analytical Paths","A new system uses Monte Carlo Tree Search to mass-produce training data for AI agents that analyze enterprise data — no human labeling required.","A research system called TOFFEE can generate large volumes of synthetic training trajectories for AI-powered data agents, targeting the specific data environments those agents will work in.\n\nData agents — LLMs wired up to query databases and run analytical workflows — tend to fall apart when dropped into unfamiliar enterprise setups. The usual fix is labeled training data showing how to handle those environments, which is slow and expensive to produce by hand. TOFFEE automates that process using Monte Carlo Tree Search, a planning algorithm borrowed from game-playing AI, to explore possible analytical paths and select the most useful ones. The system includes a task pool builder, a trajectory explorer, and a learned cost model that judges which paths are worth keeping. It also reuses common sub-sequences across tasks to cut redundancy.\n\nThe result matters because it targets both of the ways synthetic data actually gets used: fine-tuning a model on the target domain, and stuffing relevant examples into the context window of a general-purpose LLM at inference time. That dual-use design makes TOFFEE more practical than systems that optimize for only one path.\n\nMonte Carlo Tree Search has a long track record in game AI — AlphaGo leaned on it heavily — but applying it to open-ended enterprise data workflows is a messier problem with far less structured reward signals, and the paper does not yet show how well TOFFEE holds up against real-world enterprise data chaos.","[\"ai\",\"data-agents\",\"synthetic-data\",\"machine-learning\"]","2026-07-08T04:00:00.000Z","2026-07-08T04:58:02.294Z","2026-07-08T04:58:05.248Z","published",null,[],"ai",[24,26,27,28],"data-agents","synthetic-data","machine-learning",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.06233",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":86,"slug":87,"count":83,"latest_published_at":88},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]