Building a Better Q-Learning Algorithm for Reinforcement Learning

Join our event to learn the required components of a Q learning algorithm, different options for each component, and how to build each component. This will include: a) Replay buffers - for storing past data b) Fitted and gradient descent for Q learning c) Different targets for the Q function d) SARSA e) SARSA(lambda) f) Classic Q learning g) Multi-step returns Q learning h) Target networks for stabilizing our policy i) Double Q learning j) Adding CNNs. Finally, with all these tools, we will put everything together to create a more advanced Q-learning algorithms than the DQN paper.