An Introduction to Q-Learning Part 1

Hugging Face Blog · 4 years ago

Hugging Face published an educational article introducing Q-Learning as part of its free Deep Reinforcement Learning course, covering value-based methods and comparing Monte Carlo with Temporal Difference Learning approaches. The course is structured in 2 parts, with the first explaining fundamental concepts like state-value and action-value functions, and the second implementing a Q-Learning agent trained on environments including Frozen-Lake-v1 and an autonomous taxi simulator. This foundational unit is designed to prepare students for Deep Q-Learning in unit 3, which applies these principles to games like Atari breakout and space invaders.