Deep Q-Learning (DQN)

Run on Neu.ro

Explore on GitHub

The first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning was presented by Mnih et. al..

The model was a convolutional neural network, trained with a variant of Q-learning, whose input is raw pixels and whose output is a value function estimating future rewards.

Car

We apply the approach discovered in that paper to one of the traditional OpenAi gym environments - Mountain Car.

Classic DQN represents a rather simplistic approach, but at the same time, the recipe is an excellent starting point for diving into Deep Reinforcement Learning.

Quick Start

0. Sign up

1. Install CLI and log in

pip install -U neuromation
neuro login

2. Run the recipe

git clone git@github.com:neuromation/ml-recipe-mountain-car.git
cd ml-recipe-mountain-car
make setup
make jupyter