Sample code for "Tensorflow without a PhD" series.

tensorflow-rl-pong

A neural network trained to play the game of Pong for just the pixels of the game. Uses reinforcement learning and policy gradients. The approach can be generalized to other problem involving a non-differentiable step that cannot be trained using traditional supervised learning techniques.

---

Other samples from "Tensorflow without a PhD" talks will be transferred into this repository soon:

• Tensorflow and deep learning without a PhD
  video | slides | codelab | code

• Tensorflow, deep learning and recurrent neural networks, without a PhD
  video | slides | codelab | code

• Tensorflow, deep learning and modern convnets, without a PhD
  video | slides | code: soon

• Tensorflow, deep learning and modern RNN architectures, without a PhD
  video | slides | code: soon

• Batch normalisation: a new superpower
  video | slides | code

• Tensorflow and deep reinforcement learning, without a PhD
  video | slides | code

---

Disclaimer: This is not an official Google product but sample code provided for an educational purpose