Reinforcement Learning notes of MOOC by University of Alberta in Coursera

Notes taken during the lectures given by the MOOC of Reinforcement Learning by University of Alberta in Coursera.

The notes were taken from the next four courses.

• Fundamentals of Reinforcement Learning
• Sample-based Learning Methods
• Prediction and Control with Function Approximation
• A Complete Reinforcement Learning System (Capstone)

The information and images were taken from the courses and most of the text is part of the lecture.

Topics

These notes have information about the next topics:

• Markov decision process (MDP)
• Temporal Difference Learning (TD)
• Monte Carlo methods
• Sarsa
• Expected Sarsa
• Q-learning
• Policy gradients
• Dyna
• Dyna-Q
• Model-based methods
• Softmax policies
• Among others