Introduction to Reinforcement Learning in Python

Course materials I have used for a 12 lecture course on Reinforcement Learning and Game Theory.

Using environments from OpenAI Gym

Topics covered (as of early 2018):

• Markov Decision Processes.

• Dynamic Programming.

• Monte Carlo Learning.

• Temporal Difference Learning (Q-Learning, Sarsa).

• Function approximation methods / Deep Q-Learning.

• Policy approximation via black-box optimization.

• Actor-critic methods.

• Zero-sum games. Regret matching and fictitious play.

Some great resources:

• David Silver lectures.

• The classical book from Sutton and Barto, second edition now complete.

• Yandex Practical RL course.

No time for the full David Silver's course? Try:

• Introduction to Reinforcement Learning

• Deep Reinforcement Learning, Pieter Abbeel

For the applications in online advertising (retargeting) check Nicholas Le Roux talk