Syllabus
Marsland Chapter 11
Slides
Videos
The first part of the lecture is taken from our previous course INF3490, which covered exactly the same material as is syllabus here. The second part is new, and contains an extended example demonstrating in practice how a reinforcement learning algorithm operates, and highlights the difference between the two algorithms Q-learning and SARSA.
Optional Reading
From Q-Learning to Deep Q-Learning: https://towardsdatascience.com/reinforcement-learning-tutorial-part-3-basic-deep-q-learning-186164c3bf4
(Advanced) Article showing important limitations of state-of-the-art Reinforcement Learning: https://thegradient.pub/why-rl-is-flawed/