DQN in R

Simple DQN implementation in R
Original Paper: Playing Atari with Deep Reinforcement Learning

Pseudocode

UML

Gym stat

Link: https://gym.openai.com/evaluations/eval_iEDX4AWFRmu8gqoIkmlXA

Dependencies

• tensorflow
• reticulate

Files

.
├── Agent.R # Agent class is defined here
├── assets 
├── DQN-in-R.Rproj
├── main.R # the main file to run
├── Memory.R # Memory class is defined here (aka Experience Replay Memory)
├── README.md
└── train.R # training helper function