Learning Exploration Strategies with Meta-Reinforcement Learning

Explore meta-reinforcement learning strategies for efficient exploration in deep reinforcement learning with Chelsea Finn from Stanford University. Delve into the motivation behind meta-reinforcement learning, examine an example problem, and understand concepts such as posterior sampling, task-relevant information, and distribution over MDPs. Investigate the horizon of exploration and learn how to quantify task complexity in this insightful 31-minute lecture from the Simons Institute's Deep Reinforcement Learning series.

Intro
Motivation
MetaReinforcement Learning
Example Problem
posterior sampling
task relevant information
distribution over mdps
horizon of exploration
Quantifying task complexity