From Words to Actions - Foundation Models for Generalist Robot Brains

Explore the potential of Foundation models and their applications in robotics through this insightful talk by Andy Zeng, Staff Research Scientist at Google DeepMind. Delve into the challenges of scaling real robot data collection and learn about the lessons gained from this process. Discover how Foundation models and modularity might be leveraged to accelerate the development of useful robots. Gain valuable insights into the future of robotics, including the possibility of creating generalist robot brains capable of performing diverse tasks in unstructured environments. Understand the current limitations in robot data collection compared to the vast amounts of data used in training large language models. Benefit from Zeng's expertise in machine learning and robotics, backed by his impressive academic background and numerous accolades in the field.