MIT and NYU's FLEX lets robots generalize a single spoken correction into broad preferences
JafarNajafov · x · 2026-11-11
MIT CSAIL and NYU researchers present FLEX (Flexible Language-Guided Representation Editing), a robot learning system that infers general preferences from natural-language corrections instead of memorizing each rule. Told "don't spill coffee on my laptop," a robot using GPT-4o generalized the intent to "protect electronics" and avoided a nearby phone too; other tests generalized warnings into avoiding flammable objects and fragile objects. FLEX can add, generalize, and mask features rather than accumulating rules. It scored highest in all 7 simulated 2D navigation scenarios versus three baselines, and in a 12-participant study rated higher on all five measured dimensions including intent understanding. Still research-stage: main experiments were in simulation, with physical-robot demonstration data as补充 evaluation.
More from Embodied
- Lighthouse extends Obsidian welding foundation model to run on any robotic platform — bennash · 2026-11-11
- Sipeed's MaixStation Dock PCB returns, 14-layer SOM still in fab — AIFlow_ML · 2026-11-11
- NAE EU-US Frontiers of Engineering symposium explores AI-driven materials, robotics, neuromorphic computing — ProfBuehlerMIT · 2026-11-11
- Paper link for FLEX: robots generalize language corrections into preferences — JafarNajafov · 2026-11-11
- AI lets robotics research thrive in labs with no robotics experts — 4310sy · 2026-11-11
- Tektronix launches ~$500 12-bit DPO scope with AI MCP support in closed beta — IgorCarron · 2026-11-11