Deep Reinforcement Learning

This repository contains modular implementations of Deep Reinforcement Learning algorithms.

Projects based on Unity Environment

• Banana Collector: DQN agent collects yellow bananas while avoiding blue bananas.
• Reacher: DDPG agent controls double-jointed arm to reach ball.
• Tennis: MADDPG agents control rackets to play Tennis.

Dependencies

To set up your python environment to run the code in this repository, follow the instructions below.

1. Install conda

2. Create and activate a new environment with Python 3.6.

   • Linux or Mac:

   conda create --name drlnd python=3.6
   source activate drlnd

   • Windows:

   conda create --name drlnd python=3.6
   activate drlnd

3. Follow the instructions in this repository to perform a minimal install of OpenAI gym.

   • Next, install the classic control environment group by following the instructions here.
   • Then, install the box2d environment group by following the instructions here.

4. Clone the repository and install several dependencies.

git clone https://github.com/Zulkhuu/reinforcement-learning.git
cd python
pip install .

5. Create an IPython kernel for the drlnd environment.

python -m ipykernel install --user --name drlnd --display-name "drlnd"

6. Before running code in a notebook, change the kernel to match the drlnd environment by using the drop-down Kernel menu.

Future work

• Collaboration and Competition - Train a pair of agents to play soccer

Resources

Online courses

• UC Berkeley's Deep RL Bootcamp
• UC Berkeley's Deep Reinforcement Learning course
• David Silver's course on reinforcement learning
• DeepMind's Advanced Deep Learning & Reinforcement Learning course
• Udacity Deep Reinforcement Learning Nanodegree program

Textbooks

• Ian Goodfellow: The Deep Learning textbook
• Sutton & Barto: Reinforcement Learning: An Introduction
• Miguel Morales: Grokking Deep Reinforcement Learning

Papers

• Human-level control through deep reinforcement learning
• Deep Reinforcement Learning with Double Q-learning
• Dueling Network Architectures for Deep Reinforcement Learning
• Prioritized Experience Replay
• Noisy Networks for Exploration
• Rainbow: Combining Improvements in Deep Reinforcement Learning
• Distributional Reinforcement Learning with Quantile Regression
• Neural Episodic Control
• Deterministic Policy Gradient Algorithms
• Asynchronous Methods for Deep Reinforcement Learning
• Trust Region Policy Optimization
• Proximal Policy Optimization Algorithms
• Continuous control with Deep Reinforcement Learning
• Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments