Deep reinforcement learning project comparing DQN variants including baseline DQN, Double DQN, curriculum learning, and reward shaping.