At this point I think most people have some exposure to ML and AI. So rather than introduce the subject I’ll focus on second steps and beyond.

Going from Beginner to Intermediate

A lot of people ask me how they should get started in Machine Learning. While it is helpful to take a class in ML, it’s not strictly necessary at this point. Much can be learned from online books such as Jake VanderPlas’ wonderful book. That one also includes some necessary hacking skills (python).

Another essential intro is 3blue1brown’s short series on Neural Networks.

You should build a broad map of the field of machine learning. The most important continents on the map are Supervised Learning, Unsupervised Learning, and Reinforcement Learning. Read:

Summary of Three Areas

Then there’s the point when you crack these books or syllabi and realize that there’s a lot more to ML than Neural Networks… in fact, most intro courses barely spend two weeks on NNs. Sometimes less is more. And in the case of ML if it does the job, the less complicated the model the better. In this part of your education you should familiarize yourself with these models:

Supervised Learning

• Linear (and Multi Linear) Regression
• Logistic Regression
• Regularization
• k-Nearest Neighbors
• Support Vector Machines and Kernels
• Decision Trees / BDTs / Random Forests
• Gradient Boosting
• Naïve Bayes (and Bayesian Statistics)
• Gaussian Processes

Unsupervised Learning

• PCA (more classical stats than ML…)
• k-Means Clustering