An overview of robotics research at Google Deepmind.

The mission of the robotics team at Google Deepmind is to make robots useful in the real world through machine learning. We want to develop a robot generalist that can help humans in social environments. The robot learning research at Google Deepmind includes tasks such Manipulation, Navigation, Locomotion, Agility, etc. A major thrust of our research is how language can be used as a connective tissue between generative AI and robotics. In this talk, I will give an overview of the research. I will also mention the opportunities for faculty to collaborate and students to get involved with Google research.

Dr. Pannag R Sanketi is currently a Tech-Lead Manager in the Robotics team at Google Deepmind (previously called Google Brain) in Mountain View, California. He leads various projects on robot learning, including 1) X-embodied foundation models, 2) agile robots such as Table tennis and Catching, 3) infrastructure for connecting multiple robots, etc. Before joining the robotics team, he worked as a tech-lead in Android for many years, where he developed AI features for Android apps such as Camera, Photos and Pixel watch, founded on-device ML projects (e.g. frameworks for 3P apps such as MLKit and for 1P apps for privacy sensitive training). Prior to Google, he worked at a computer vision based startup called PixBlitz and then at the Smith-Kettlewell Eye Research Institute in San Francisco as a research associate, developing Computer Vision based cell phone algorithms to help visually impaired people. He did his undergraduate from IIT-Madras in Mechanical Engineering from 1999-2003. He obtained his MS and PhD from University of California, Berkeley in 2005 and 2009 respectively in Control Systems and Optimal Control. His research interests lie at the intersection of robotics, optimal control and machine learning.

Abstract

Dr. Pannag R. Sanketi, Google Deepmind