Adam Merchant, Ryan Strauss
Faculty Sponsor: Raghu Ramanujan
We set out to replicate the results of Mnih et. al.’s “Human-level control through deep reinforcement learning”, in which they introduced the Deep Q Learning agent (DQN). Their agent was able to master a number of classic Atari games using only high-level sensory input. We were able to successfully implement the DQN agent and have it master a number of Atari games.