Better representations first: where does data scaling break down in robot learning?
KyleMorgenstein · x · 2026-11-17
In this thread, the author contrasts two approaches to robot contact-interaction problems: the mainstream route of synthetic data generation over a huge object distribution, hoping all real interactions are covered and the policy interpolates; versus work that finds a better representation first, reducing dependence on data volume. The open question: to what extent can robot learning collect its way out of these problems, and where does that break down?
More from Embodied
- Enactra debuts embodied-AI simulation platform where every frame ships annotated — Lianhuiq · 2026-11-17
- LG targets Q1 2027 to unveil bipedal humanoid on NVIDIA reference design — xmercury_one · 2026-11-17
- Robotics researcher: edge cases are literally the entire problem for robot deployment — chris_j_paxton · 2026-11-17
- Policies use only a sliver of image information — extracting it matters for safety — KyleMorgenstein · 2026-11-17
- South Korea's robot density hits 1,066 per 10k workers, 7x the world average — randal_olson · 2026-11-17
- Open-source Human Models MCP connects AI agents to MuJoCo musculoskeletal simulation — usefulwrong · 2026-11-17