Deep Reinforcement Learning Agents

This repository contains a collection of reinforcement learning algorithms written in Tensorflow. The ipython notebook here were written to go along with a still-underway tutorial series I have been publishing on Medium. If you are new to reinforcement learning, I recommend reading the accompanying post for each algorithm.

The repository currently contains the following algorithms:

• Q-Table - An implementation of Q-learning using tables to solve a stochastic environment problem.

• Q-Network - A neural network implementation of Q-Learning to solve the same environment as in Q-Table.

• Simple-Policy - An implementation of policy gradient method for stateless environments such as n-armed bandit problems.

• Contextual-Policy - An implementation of policy gradient method for stateful environments such as contextual bandit problems.

• Policy-Network - An implementation of a neural network policy-gradient agent that solves full RL problems with states and delayed rewards, and two opposite actions (ie. CartPole or Pong).

• Vanilla-Policy - An implementation of a neural network vanilla-policy-gradient agent that solves full RL problems with states, delayed rewards, and an arbitrary number of actions.

• Model-Network - An addition to the Policy-Network algorithm which includes a separate network which models the environment dynamics.

• Double-Dueling-DQN - An implementation of a Deep-Q Network with the Double DQN and Dueling DQN additions to improve stability and performance.

• Deep-Recurrent-Q-Network - An implementation of a Deep Recurrent Q-Network which can solve reinforcement learning problems involving partial observability.

• Q-Exploration - An implementation of DQN containing multiple action-selection strategies for exploration. Strategies include: greedy, random, e-greedy, Boltzmann, and Bayesian Dropout.

• A3C-Doom - An implementation of Asynchronous Advantage Actor-Critic (A3C) algorithm. It utilizes multiple agents to collectively improve a policy. This implementation can solve RL problems in 3D environments such as VizDoom challenges.