October 30, 2015

In part 3 of the reinforcement learning series we implement a neural network as the action-value function and use the Q-learning algorithm to train an agent how to play Gridworld.

Read More
October 25, 2015

Part 2 of the RL series. A slightly deeper dive into reinforcement learning methods by learning how to use Monte Carlo simulations to learn how to play blackjack.

Read More
October 19, 2015

The first part in a series introducing the theory, math and implementation details of reinforcement learning algorithms using Python. Here we introduce the topic with a very simple RL problem, the n-armed bandit problem.

Read More
September 29, 2015

A gentle introduction to the powerful machine learning library, Theano.

Read More
September 12, 2015

A brief addendum to my previous post on a simple genetic algorithm. Here I explore how varying the parameters affects our GA performance.

Read More