ML & AI: Machine Learning for Researchers
Machine learning is easier to learn once you know regression. A neural network is stacked logistic regressions, and a random forest is a committee of if-then rules. This course is for you if you know the Statistics Core and want the ML vocabulary without the hype or the heavy math.
The main difference is the goal. Classical statistics asks why; machine learning asks how well can we predict the next case. That change affects everything else, starting with why you must hold out a test set and never grade a model on data it has seen. The course then covers regularization for models with many predictors, the accuracy trap and ROC curves, tree-based models (trees and forests), clustering and dimensionality reduction for data with no labels, and what neural networks and large language models can and cannot do in a research workflow.
The lessons assume you know regression at the level of Stats 3, and they refer back to it often.