I am currently working on developing a new AI programming course at Master’s level.
What is it about?
In this course, I explore an often-overlooked aspect of AI: Reinforcement learning. In addition, a particular chapter about pathfinding and its applications in navigation systems is planned.
I have the chance to have complete autonomy over the entire 35-hour module. The classes are divided between fundamental reinforcement learning theories (value-based learning, policy-based learning, Monte Carlo methods, etc.) and an exciting project in which students can train an autonomous car agent on a track using the Unity game engine.
It was a lot of work to create the whole project, but the insights and takeaways for the students were worth it.
I am almost halfway through the unit, and I am impressed so far by the students’ interest in the subject, even wanting to go deeper into the mathematical theorems involved!