NVIDIA's AlphaNIST Learns Robotic Assembly From Scratch With Only Binary Rewards, Hitting 98.2% Across 20 Tasks
abhishekunique7 · x · 2026-10-14
NVIDIA Research's AlphaNIST paper, inspired by AlphaGo Zero, trains a robot from tabula rasa using only a sparse binary task reward plus a simple control cost — no demonstrations, progress shaping, scripted primitives, or per-task tuning. Key results:
- A single policy trained from scratch in simulation solves all 20 assembly tasks on the NIST Assembly Task Board #1 at 98.2% overall success
- Policies also complete 10 different assembly pairs in one episode with no prescribed order, rewarded only when all ten pairs are simultaneously complete
- Emergent behaviors include dexterous grasp selection and planning-like strategies
Paper: arxiv.org/abs/2610.13354
More from Embodied
- China reportedly shipped 97% of the world's humanoid robots in H1 2026 — Polymarket · 2026-10-14
- Physical Intelligence opens Pi API beta, one model adapting robots from strawberry harvesting to server assembly — svlevine · 2026-10-14
- China's Tiangong Ultra robot beat Bolt's 100m record — and exposes America's robotics supply-chain dilemma — mattsheehan88 · 2026-10-14
- GoDrift shows off realtime sim2real robot demo; 2D plane eases reaction demands — jonstephens85 · 2026-10-14
- Microsoft science president Peter Lee leaves to become Mercedes F1 team's AI advisor — lazowska · 2026-10-14
- NVIDIA releases GR00T N1.7 AppleToPlate checkpoint-15000 on Hugging Face — nvidia · 2026-10-14