Samory Kpotufe

Unsupervised Learning algos, regression, lasso, ridge, non-parametric kernel methods, classification, theory in ML.

kernel machine, RKHS stuff. nonparametric minimax.

basic probability, types of variables, continuous and discrete variables, estimation, confidence intervals, hypothesis Testing, regression

Traditional nonparametric methods such as k-NN (classification and regression), kernel density estimation. Free lunch theorem and how smoothness of the function space can give rise to generalization bounds and convergence rates. Kernel machines, Hilbert spaces, RKHS, infinite dimensional function spaces.

This course was outstanding. Having taken a previous introductory AI course, this one delved deeply into the mathematical models and intuitions behind common ML models, including k-NN, various classifiers, neural networks, SVMs, Gaussian mixtures, and more. It certainly demands a solid math background to keep up; otherwise, you'll need to invest a lot of time and effort to get up to speed.

Some basic and advanced ML algorithms + PyTorch.

Probability, random variables, hypothesis testing, linear approximation, PDFs and CDFs, etc.

Non-parametric methods, no free lunch theorem, classification is easier than regression, assumptions under which minimax rates improve, abstract treatment of linear operators, Hilbert spaces, techniques to complete a space, RKHS.

The class is much smaller than the COMS ML course, which lends itself better to asking more questions, and allows the prof to go at a better pace. more rigorous and theoretical than the COMS course and overall i think it was a solid introduction to machine learning topics.

I learned about topics such as different kinds of distributions and how to conduct tests and how to analyze test statistics.