My forthcoming book Causal Machine Learning is available on early access! Use the code mlness to get 40% off until August 30th.

Sean J. Taylor on Twitter: "Very excited about @osazuwa’s new book. I haven’t had a chance to read yet, but the topics he covers are sorely underrepresented in how most of us learn both causal inference and ML."

Why I’m writing this book

I wanted to write a code-first introduction to causal ML, as code was always the best way for me to learn new ideas.

Also, I wanted to write the book because no one else was making the (at least to me) obvious connections between causality and probabilistic machine learning. I was mystified by this missing link, given so many recent advances in the latter area, particular in probabilistic deep learning.

For example, I was reading through a recent book called Causal Inference the Mixtape. Good book! Would recommend. But the author has these detailed mathematical derivations that I, with my fancy PhD in stats, still had to reread a few times, recalling stuff that only people with graduate degrees in statistics, economics, or a related field would reasonably be able to recall.

It’s a common theme in causality books. And as I read about one such method, I was thinking how easy and intuitive it would be to implement the model using variational inference in Pytorch.

Even if a reader didn’t retain all the math, at least they’d have gotten better at Pytorch - a skill more marketable than learning how to run R libraries with opaque code written by academics!

So that’s what I’m trying to accomplish with this book. If you’re interested, I hope you’ll check out the early access version. Please feel free to reach out with your feedback. The book is a work in progress and it’s likely I’ll incorporate your feedback into the final version.