Introduction
This is the last part of a four parts long series about reinforcement learning. In this part an interesting case study is covered: the replication of the Double DQN algorithm.
Case Study: Double Deep Q-Network
This is a short post because every detail is available on github. As a short summary it is about the Double Deep Q-Network algorithm which learns to play the game Breakout in the Atari environment. The implementation is in Keras (with Tensorflow backend) and OpenAI gym provides the Atari emulator to interact with the game. For further information follow the link below.