Learning in Robotics
The discussion explores the potential of solving robotics challenges through end-to-end learning rather than traditional methods that rely on human expertise. Sergey emphasizes that while the terminology may have evolved, the core idea of automated optimization has been around for a long time. A significant distinction lies in the ability of learning-based systems to improve continuously through experience, contrasting with classic optimal control systems.In this clip
From this podcast
Lex Fridman Podcast
Sergey Levine: Robotics and Machine Learning | Lex Fridman Podcast #108
Related Questions
What does Lex Fridman say about programming and learning in the episode Sergey Levine: Robotics and Machine Learning | Lex Fridman Podcast #108 and the clip Learning in Robotics?
Are there limits to human knowledge as discussed in the episode Sergey Levine: Robotics and Machine Learning | Lex Fridman Podcast #108 and the clip Learning from Data?