Sim-humanoid learns basketball in IsaacGym via soft-routed skill primitives
ericjang11 · x · 2026-10-15
New sim research, "Learning to Ball," trains a physics-simulated humanoid to play basketball inside IsaacGym.
- Seven primitive skills (dribble, gather, shoot, pass, catch, rebound, defend/locomotion) are learned via adversarial imitation, then composed into long-horizon plays
- The approach replaces hard skill-chaining and one-hot mixture-of-experts, which break when subtask policies share few states
- The core idea is a soft router that blends all primitives every timestep; disabling it sharply drops shooting accuracy and it beats the one-hot hard-switching baseline
More from Embodied
- Gemini controls a real robot from one prompt: SSH in, find joints, learn from demo — heiga_zen · 2026-10-15
- Tesla adds 30 more Cybercabs in Texas, purpose-built fleet hits 349 with 300+ serving Austin — XFreeze · 2026-10-15
- Physical AI webinar Oct 22: four robotics founders on evals, retraining, and field failures — carrycooldude · 2026-10-15
- Vela: spline-based continuous trajectories scale vision-language-action models — Yifan Li · 2026-10-15
- Scoble: healthcare robots arrive next year, true humanoids likely 2-7 years away — realmeetjames · 2026-10-15
- KAIST's BeetleBot: finger-controlled soft robot walks, grips and streams sensory feedback — CurieuxExplorer · 2026-10-15