First lecture on August 22

CS294-112

CS 294-112 at UC Berkeley

Deep Reinforcement Learning

Lectures: Wed/Fri 10-11:30 a.m., Soda Hall, Room 306


The lectures will be streamed and recorded. The course is not being offered as an online course, and the videos are provided only for your personal informational and entertainment purposes. They are not part of any course requirement or degree-bearing university program.


If you are a UC Berkeley undergraduate student or non-EECS graduate student and want to enroll in the course for fall 2018, please fill out this application form. We will select students from this list in August based on space availability and prerequisites.

  • Instructor Sergey Levine

    svlevine@eecs.berkeley.edu

    Office Hours: Wed 11:30am - 12:30pm (Soda 341B)

  • Head GSI Kate Rakelly

    rakelly@eecs.berkeley.edu

    Office Hours: Tue 11am - 12pm (Soda 651)

  • GSI Gregory Kahn

    gkahn@berkeley.edu

    Office Hours: Thu 1pm - 2pm (Soda 411)

  • GSI Sid Reddy

    sgr@berkeley.edu

    Office Hours: Fri 3pm - 4pm (Soda 411)

  • GSI Michael Chang

    mbchang@berkeley.edu

    Office Hours: Mon 5pm - 6pm (Soda 611)

  • uGSI Soroush Nasiriany

    snasiriany@berkeley.edu

Week 1 Overview

Course Introduction, Imitation Learning

Homeworks

See Syllabus for more information.

  • Homework 1: Imitation Learning
  • Homework 2: Policy Gradients
  • Homework 3: Q-Learning and Actor-Critic
  • Homework 4: Model-Based RL
  • Homework 5: Advanced Topics
Expand

Lecture Slides

See Syllabus for more information.

  • Slide 1: Introduction and Course Overview
  • Slide 2: Supervised Learning and Imitation
  • Slide 3: TensorFlow and Neural Nets Review Session
  • Slide 4: Reinforcement Learning Introduction
  • Slide 5: Policy Gradients Introduction
  • Slide 6: Actor-Critic Introduction
  • Slide 7: Advanced Q-Learning Algorithms
  • Slide 8: Advanced Actor-Critic Algorithms
  • Slide 9: Advanced Policy Gradients
  • Slide 10: Optimal Control and Planning
  • Slide 11: Learning Policies by Imitating Optimal Controllers
  • Slide 12: Learning Dynamical Systems from Data
  • Slide 13: Advanced Model Learning and Images
  • Slide 14: Probability and Variational Inference Primer
  • Slide 15: Connection between Inference and Control
  • Slide 16: Inverse Reinforcement Learning
  • Slide 17: Exploration: Part 1
  • Slide 18: Exploration: Part 2
  • Slide 19: Transfer Learning and Multi-Task Learning
  • Slide 20: Meta-Learning
  • Slide 21: Parallelism and RL System Design
  • Slide 22: Advanced Imitation Learning and Open Problems
  • Slide 23: Guest Lecture 1
  • Slide 24: Guest Lecture 2
  • Slide 25: Guest Lecture 3
  • Slide 26: Guest Lecture 4
  • Slide 27: Final Project Presentations: Part 1
  • Slide 28: Final Project Presentations: Part 2
Expand