Martin Ferianc

PhD at UCL
PhD student at @UCL, @IIMLUCL interested in machine learning, Bayesian neural networks, ballroom dancing 🕺 and cats 🐈.

Introduction to Reinforcement Learning

Hello Everyone,

Together with Xiaoliang Luo we wrote a comprehensive tutorial to introduce eager readers to a high-level overview of the fundamentals of reinforcement learning as well as a Python example code, working with the OpenAI Gym library. More specifically, we teach an agent, using Q-learning implemented from scratch, to balance a pole on a cart. Check it out here:

https://deepnote.com/project/Intro-to-Q-learning-in-RL-4R450s6_RVKJIC2xiqs71g/%2Frl_qlearning.ipynb


Goals of the tutorial:
Introduce you to:
  • Reinforcement learning foundations not omitting the maths
  • Programming reinforcement learning algorithms (Q-learning) from scratch
  • Interacting with OpenAI gym library
  • Visualisation of your agent playing the game

Outline:
  • Reinforcement learning motivation along with an example
  • Fundamental notation in reinforcement learning
  • Q-learning
  • OpenAI gym
  • Programming and using Q-learning to train an agent
  • Final test and visualisation

Prerequisites:
  • Python programming experience (beginner-intermediate)
  • Foundations of statistics and probability

Submitted as a part of Deepnote publishing competition

Like Comment