Reinforcement Learning Insights
Daniel and Sergey discuss a recent paper that explores reinforcement learning as a sequence modeling problem. They delve into the concept of tokenizing trajectories and training a large sequence model using a transformer. The paper's approach, intentionally made to be as obtuse as possible, offers valuable insights into alternative ways of solving problems in the field of reinforcement learning.In this clip
From this podcast
The Gradient
Sergey Levine on Robot Learning & Offline RL
Related Questions