RL problems are canonically implemented as a Markov Decision Process. The MDP defines the dynamics of the environment. Some of these dynamics are easy to see and understand while others less so. Visualizing certain attributes of the domain can significantly impact user’s ability to understand and interpret an agent’s behavior.
This project will aim to provide visualizations for some of the following attributes:
1.Agent Rewards – Built in rewards of the Domain
2.Agent Uncertainty – States where the agent is less confident regarding its choice of actions
3.Pareto-Landmark States – Areas of the environment which must be part of any solution.