Module Overview
This course teaches students the fundamentals of reinforcement learning and its elements. As part of the course, students would be introduced to OpenAI gym - which is a programming environment used for implementing RL agents. The key objective being to familiarize students with basic RL algorithms and applications. The emphasis would thus be on algorithms and applications, with some broad explanation of the underlying principles.
Learning Outcomes
Upon successfully completing the course, you will be able to:
Explain the following concepts:
Episodic and continuing tasks
Reward hypothesis, goals and rewards, cumulative rewards and discounted returns
Markov decision processes (MDP)
Policy and value functions (state value function, action value function)
Implement and train a Deep Q network
Implement and train a policy gradient network