Advanced Reinforcement Learning Course

Welcome to the repository for the CS 7180 special topics in AI course, focusing on Advanced Reinforcement Learning for Ph.D. students. This comprehensive course covers a wide range of topics, including fundamental concepts such as Markov Decision Processes (MDP), Dynamic Programming, Temporal-Difference Learning, Q-learning, SARSA, and Monte-Carlo Tree Search (MCTS).

As the course progresses, we delve into more advanced techniques, exploring Deep Reinforcement Learning (Deep RL) with algorithms like Deep Q-Network (DQN), Policy Gradient (PG), and actor-critic methods such as REINFORCE, DDPG, and PPO. Our journey extends beyond the basics to cover Partially Observable RL, Multi-agent RL, and Representation Learning within the context of Reinforcement Learning. Furthermore, we explore the practical application of RL in the field of Robotics.

Course Details

• Course Title: Advanced Reinforcement Learning (CS 7180 Special Topics in AI)
• Exercises:
  1. Exercise 0: An Invitation to Reinforcement Learning
  2. Exercise 1: Multi-armed Bandits
  3. Exercise 2: Markov Decision Processes (MDPs)
  4. Exercise 3: Dynamic Programming
  5. Exercise 4: Monte-Carlo Methods
  6. Exercise 5: Temporal-Difference Learning
  7. Exercise 6: Planning and Learning
  8. Exercise 7: Function Approximation
  9. Exercise 8: Deep Q-Network (DQN)
  10. Exercise 9: Policy Gradient (PG)

The course structure includes 10 exercises, each designed to build a strong foundation in reinforcement learning principles, followed by a research-based project. Feel free to explore the exercises and enhance your understanding of advanced reinforcement learning techniques.