Some free beginner courses from Google:

Introduction to Generative AI Learning Path

This learning path provides an overview of generative AI concepts, from the fundamentals of large language models to responsible AI principles.

01 Introduction to Generative AI
02 Introduction to Large Language Models
03 Introduction to Responsible AI
04 Prompt Design in Vertex AI
05 Responsible AI: Applying AI Principles with Google Cloud

Introduction to Statistical Learning by James, Witten, Hastie, and Tibshirani is my highest recommended intro to machine learning text:

• https://www.statlearning.com/

It’s where I think most folks should start if they have the requisite math background. It currently has R and Python versions (I used the R version). It was actually a goal of mine at one point to write an Elixir version, but I’ve never come back to it

I also think Deep Learning by Goodfellow, Bengio, Courville is excellent if you’re more interested in Neural Networks, specifically:

• https://www.deeplearningbook.org/contents/intro.html

Hugging Face courses: Hugging Face - Learn

It’s also worth checking out their (community) blog here: Hugging Face – Blog
You can search for topics and they have everything from basics to advanced. As you could expect some articles are of higher quality than others.

By far I feel the one that has helped me the most is this course https://www.coursera.org/specializations/machine-learning-introduction

It’s by Andrew Ng, and takes you through the math behind machine learning, step by step. I believe understanding the mathematics and why these algorithms exist is important for truly understanding machine learning.

What I did to re-enforce my learning is I would write out the code examples in Elixir using Nx.

Since I’m more comfortable in Elixir than in Python. Along the way I also get to learn and understand the different parts of Nx and what all the functions are used for.

I’ve really been enjoying this process. Even though one can simply use Scholar to do things like linear regression and logistic regression it’s still good to write out the various components so I can understand what’s going on behind those function calls.

I went from “calculus ewwww…” to “wow calculus is awesome” in the span of about 3 weeks. This was the course that made me change.

After understanding things a bit better I went back to read Machine Learning in Elixir Machine Learning in Elixir: Learning to Learn with Nx and Axon by Sean Moriarity and things started to click much much better.

Oh that reminded me of the fastai course I followed some time ago.
It’s a series of YouTube videos and accompanying notebooks.

Here are a few articles:

• Simplest Implementation of Diffusion Models

• Differentiable Programming from Scratch

• Exploring Neural Graphics Primitives

• Why Train When You Can Optimize?

• Solving Transformer by Hand: A Step-by-Step Math Example

• Hacker’s guide to Neural Networks

(I kind of go for the “from scratch” manner of exposition.)

Two slightly older books are pretty good, cause they cover a lot of ground, and well-understood “greatest hits” kind of problems:

• Programming Collective Intelligence
• Elegant SciPy

I really like Neural Networks: Zero To Hero by Andrej Karpathy. I’ve recently picked up Elixir in the last few months and have been going through the series and following along using Livebook + Nx ecosystem.

I wish there were more Elixir specific books. There is only one in Pragmatic Bookshelf, and none in Manning Publications.

It’s only a matter of time

Here’s a couple for you to get on with

Thanks!