Success-Guided Sampling: sim-to-real RL nails dexterous assembly with zero demos
kevin_zakka · x · 2026-10-10
UW and NVIDIA researchers (CoRL 2026) release Success-Guided Sampling (SGS), a simple tweak to the task sampler that fixes the exploration bottleneck in mega-scale RL. Policies trained purely in simulation—with no demos, no tactile sensors, no per-task reward tuning—mesh gears, thread nuts, and insert pegs zero-shot on real UR5e/Franka arms from RGB, and a single policy handles ANYmal C/D across extreme terrains. Scales to 1M+ parallel simulated robots using only PPO. Code coming soon.
More from Embodied
- Artificial Analysis teases AA-Robotics: frontier models zero-shot robot arm control — ArtificialAnlys · 2026-10-10
- Success-Guided Sampling: zero real-world data for dexterous robot hands — milesmacklin · 2026-10-10
- Starkey Unveils Omega AI+ Hearing Aids With G4 Gen AI Processor and Quad DNN — BrandonSawalich · 2026-10-10
- Wayve CEO demos AI Driver with Stellantis in a Fiat 500e on Turin's tight streets — alexgkendall · 2026-10-10
- 192GB Framework Desktop Batch 2 sells out; 128GB still in stock — film_girl · 2026-10-10
- Danu Robotics' six-year fight to build a better recycling robot — TechCrunch AI · 2026-10-10