Deep Q-Network & Dueling network architectures for deep reinforcement learning