CS 747 - Foundations of Intelligent Learning Agents

Programming Assignments

1. Implemented various algorithms - Epsilon Greedy, Round Robin, UCB, KL-UCB and Thompson Sampling and compared the regrets over different horizons.
2. Implemented Linear Programming solver and Howard's Policy Iteration to find the optimal policy and the corresponding value functions.
3. Estimated the Value Function for different states using Model-Based and TD(lambda).
4. Used SARSA On-Policy TD Control method to train an agent to reach the goal block of a windy gridworld. (Sutton and Barto Example 6.5, Exercise 6.9, Exercise 6.10)