Ambient Diffusion Wins Dual Awards for Training Robots with Suboptimal Data

The paper "Ambient Diffusion Policy" won dual awards at the RSS 2026 conference. It introduces a novel method that enables robots to learn from suboptimal data by modifying just one line of code, significantly improving imitation learning performance.

2026-07-30 ~ 2026-07-30 · 2 related posts

• Ambient Diffusion Policy Wins Two Best Paper Awards at RSS 2026 — giannis_daras · 2026-07-30
• Extracting Learning Signal from Suboptimal Robot Data with Ambient Diffusion — giannis_daras · 2026-07-30