CS 294: Deep Reinforcement Learning (Spring 2017, UC Berkeley) | Computer Science

CS 294: Deep Reinforcement Learning

CS 294: Deep Reinforcement Learning (Spring 2017, UC Berkeley). Instructors: Sergey Levine, John Schulman, and Chelsea Finn. This course will assume some familiarity with reinforcement learning, numerical optimization and machine learning. The course covers topics: Supervised learning and decision making; Basic reinforcement learning: Q-learning and policy gradients; Advanced model learning and prediction; Advanced deep reinforcement learning: trust region policy gradients, actor-critic methods, exploration; Open problems and research talks.

Lecture 01 - Introduction

Lecture 02 - Supervised Learning of Behaviors: Deep Learning, Dynamical Systems, and Behavior Cloning

Lecture 03 - Optimal Control, Trajectory, Optimization, and Planning

Lecture 04 - Learning Dynamical System Models from Data

Lecture 05 - Learning Policies by Imitating Optimal Control

Lecture 06 - Direct Collocation Methods for Trajectory Optimization and Policy Learning

Lecture 07 - Markov Decision Processes and Solving Finite Problems

Lecture 08 - Policy Gradient Methods

Lecture 09 - Q-Function Learning Methods

Lecture 10 - Advanced Q-Function Learning Methods

Lecture 11 - Advanced Model Learning

Lecture 12 - Advanced Topics in Imitation Learning and Safety

Lecture 13 - Inverse Reinforcement Learning

Lecture 14 - Advanced Policy Gradient Methods: Natural Gradient, TRPO, and More

Lecture 15 - Variance Reduction for Policy Gradient Methods

Lecture 16 - Policy Gradient Methods: Pathwise Derivative Methods and Wrap-Up

Lecture 17 - The Exploration Problem

Lecture 18 - Asynchronous and Parallel Algorithms

Lecture 19 - Transfer in (Deep) Reinforcement Learning

Lecture 20 - Neural Architecture Search with Reinforcement Learning

Lecture 21 - Generalization and Safety in Reinforcement Learning and Control

Lecture 22 - Deep Reinforcement Learning with Forward Prediction, Memory, and Hierarchy

Lecture 23 - Towards a Unified View of Supervised Learning and Reinforcement Learning

Lecture 24 - Adversarial Examples in Reinforcement Learning

Lecture 25 - Review

References

CS 294: Deep Reinforcement Learning (Spring 2017)
Instructors: Sergey Levine, John Schulman, and Chelsea Finn. Lecture Slides. Readings. Assignments. This course will assume some familiarity with reinforcement learning, numerical optimization and machine learning.