This is an introduction to Reinforcement Learning using Q-Learning and Deep Q-Learning.
Q Learning is demonstrated using OpenAI's Frozen Lake Environment. You can read more about the environment Here.
- Markov Decision Process and it's components i.e. The Agent, Environment, State, Action, Reward
- Expected Discounted Return (Discount Rate)
- Epsilon-Greedy Strategy
- Bellman Optimality Equation (Optimal policy, Optimal Q function, Learning Rate)
Deep Q Network is demonstrated using OpenAI's Cart Pole Environment. You can read more about the environment Here.
- Artificial Neural Network Architecture
- Experience Replay
- Training and Target Networks
P.S. Both codes are well commented/documented for ease of understanding.