Policy Representation Trade-offs
Leslie discusses the balance between different approaches in reinforcement learning, highlighting the trade-offs between model-based and model-free methods. She emphasizes the importance of representation in policy learning, suggesting that the choice of model can impact both the efficiency of computation and the speed of action selection. The conversation also touches on the varying requirements of different tasks, such as algebra manipulations versus controlling a unicycle, illustrating the complexity of decision-making in dynamic environments.In this clip
From this podcast
Lex Fridman Podcast
Leslie Kaelbling: Reinforcement Learning, Planning, and Robotics | Lex Fridman Podcast #15
Related Questions