Date: Wednesday, November 5, 2025
Speaker: Karl Pertsch, Technical Staff at Physical Intelligence
Title: Building and Evaluating Generalist Robot Policies
Abstract: In this talk, I will give an overview of modern approaches for training generalist robot policies. I will cover the design space of vision-language-action models (VLAs), that underpin most recent robot manipulation results like Pi0 and Pi0.5, and explain the modeling advances we developed over the last 1.5 years that made these possible. I will also discuss our work on scalable benchmarking approaches for such policies as a step towards community-wide real-robot benchmarks.
Bio: Karl Pertsch is a member of the technical staff at Physical Intelligence. Before, he was a postdoc at UC Berkeley and Stanford, working with Sergey Levine and Chelsea Finn. His work focuses on building generalist robot policies that can solve a wide range of physical manipulation tasks in the real world. Karl obtained his PhD from USC, advised by Joseph Lim. During his PhD he interned at MetaAI and Google Brain. His work has been awarded the Best Conference Paper Award at ICRA'24, two Outstanding Paper Awards Finalists at CoRL'24, and a Best Paper Finalist at RSS'25.
