Machine Learning in TensorFlow

The purpose of this repository is to offer anyone who wants to get started with using TensorFlow a number of experiments, where the same problem is solved through different techniques, showcasing different methods of approaching the same problem.

The implementations and parameters aren't necessarily optimized to solve the problem, since that isn't the purpose, but rather serve as a solid starting ground to play around with for your own experiments and optimizations!

It is currently under construction, but if you have any feedback, suggestions or comments - I'd gladly hear it!

Reinforcement Learning

Implemented Examples

• Q-Gridworld
• Classic Control with Policy Gradient (Work in progress)
• Deep Q-Learning in Atari (Work in progress)

Good reads

• Deep Reinforcement Learning: Pong from pixels
• Dissecting Reinforcement Learning - (Part 1, Part 2, Part 3)
• Demystifying Deep Reinforcement Learning - (Part 1, Part 2, Part 3)
• Asynchronous Deep Reinforcement Learning from pixels

Neural Networks

Implemented Examples

• XOR (Work in progress)
• NN-MNIST
• NN-Gridworld

Good Reads

• Machine Learning Crash Course - (Part 1, Part 2)
• Neural Networks (part of CS231 at Stanford) - (Part 1, Part 2, Part 3)
• Neural Networks and Deep Learning

Recurrent Neural Networks

Implemented Examples

• RNN-MNIST

Good reads

• The Unreasonable Effectiveness on Recurrent Neural Networks
• Understanding LSTM Networks

Convolutional Neural Networks

Implemented Examples

• CNN-MNIST

Good Reads

• Convolutional Neural Networks (part of CS231 at Stanford)
• How do Convolutional Neural Networks work?