Reinforcement Learning - Pac-Man

Explore reinforcement learning through a captivating conference talk that demonstrates how to build an agent capable of beating Pac-Man. Delve into both traditional Q-learning and advanced deep Q-learning techniques utilizing neural networks. Gain insights into reinforcement learning theory, learn to construct deep learning models using libraries like Keras, and witness the development of a Pac-Man-beating agent. While prior machine learning knowledge is beneficial, it's not required to appreciate this hour-long presentation from NDC Oslo 2020. Watch as speakers Malte Loller-Andersen and Manu Gopinathan explain the techniques and libraries employed, showcasing the agent's learning process along the way.