ETH Zurich Introduces NaP-Control: Merging RL with Diffusion Priors for Fast Character Control
rsasaki0109 · x · 2026-08-23
ETH Zurich presents NaP-Control, a latent noise optimization framework combining reinforcement learning (RL) with a task-agnostic diffusion motion prior. Instead of relying on slow, iterative gradient guidance at test-time, it learns to navigate the input noise space of the diffusion prior to generate task-driven, physically plausible behaviors, achieving efficient and precise whole-body character control.
More from Embodied
- 2nd World Humanoid Robot Games kicks off in Beijing with quadrupled robot count — 2C_ornot2C · 2026-08-23
- OakBot releases cooking multimodal model and autonomous kitchen, using real-world data flywheel — 新智元 · 2026-08-23
- Robot Leg Update: Switch to Aluminum for Predictable Manufacturing — TinfoilTricorn · 2026-08-23
- After beating human running speed, robot now breaks standing high jump record — iamfakhrealam · 2026-08-23
- Video: Robot scores a KO at world humanoid robot fight — iamfakhrealam · 2026-08-23
- Real-time Lunar Landing Site Detection Pipeline Using YOLOv8 and Kalman Filtering — rsasaki0109 · 2026-08-23