Central AI

Welcome to Central AI, an educational website to sharpen your machine learning knowledge!

Background

Hi, my name is Jordan and I am the author of Central AI, a resource that aims to centralise knowledge on Data Science - specifically Machine Learning and Reinfocement Learning. After completing my undergraduate degree in Actuarial Sciences at the University of East Anglia in 2015, I worked under the asset management department within NHS Property Services as an analyst until late 2017. I then completed my MSc Data Science at the University of Bath and decided to pursue a PhD. Since then I have completed an MRes in Statistical Applied Mathematics and am currently on the Statistical Applied Mathematics Bath (SAMBa) CDT. I have found during my time as a Data Science student, and now tutoring it, that easy to access online material for a lot of machine learning concepts are not well covered nor centralised. This has motivated me to attempt to fill in the gaps with tutorials, where I implement these concepts and algorithms in Python, all in one central location.

I have a few tutorial channels, each with a different theme or purpose.

How to: Machine Learning

A beginner’s guide to machine learning designed for those that know a little Python and some key terms. Suited for those in education who want to understand the algorithms. A from scratch attitude is adopted here whereby most things will be built using numpy and scipy instead of importing off-the-shelf algorithms from sklearn (scikit-learn).

How to: Reinforcement Learning

An introduction to Reinforcement Learning starting from what it is, and going from the basics with Markov Decision Processes to function approximation with Neural Networks. Tabular methods are implemented in numpy whilst function approximation is in PyTorch.

Advanced Applications

Building further on the How to tutorials series, apply ML and RL techniques on more interesting and more realistic problems. Most implementations will be in PyTorch.

Theories

A bit more math-heavy, this set of tutorials looks to answer the theoretical problems present in machine learning.