A Deep-Q Network (DQN) is a reinforcement learning technique that attempts to model the actions that perform best in each state in real-time.