SuperMimic wins CoRL Best Paper for differentiable world model
According to Berkeley AI Research citing CoRL, the paper SuperMimic: Scaling Humanoid Whole-Body Control with One Differentiable World Model won Best Conference Paper in Austin.
SourceAnalysis
Researchers from the labs of Roberto Horowitz, Pieter Abbeel and Koushil Sreenath at Berkeley AI Research received the Best Conference Paper award for their work SuperMimic: Scaling Humanoid Whole-Body Control with One Differentiable World Model at the Conference on Robot Learning in Austin, according to an announcement by Berkeley AI Research on November 12, 2026.
The award was presented yesterday at the conference, with additional confirmation from lead author Qiayuan Liao, see the Conference on Robot Learning awards page for details.
Berkeley AI Research
@berkeley_aiWe're graduate students, postdocs, faculty and scientists at the cutting edge of artificial intelligence research.