Statistical Machine Learning

via Independent

Overview

Embark on a comprehensive journey through the world of Statistical Machine Learning with this advanced second graduate level course. Tailored for students who have previously completed courses in Machine Learning (10-715) and Intermediate Statistics (36-705), this program emphasizes statistical theory and methodologies within the machine learning landscape. It expertly blends theoretical foundations with practical methodology, offering insights into selecting the most suitable methods and approaches for research problems.

The course curriculum delves into critical topics of statistical theory pivotal for machine learning researchers, including nonparametric theory, consistency, minimax estimation, and the concentration of measure. It begins with a review of essential concepts such as probability, bias/variance analysis, maximum likelihood estimation (mle), regression, and classification before advancing to the theoretical foundations covering Function Spaces (including Holder spaces, Sobolev spaces, and reproducing kernel Hilbert spaces (RKHS)), Concentration of Measure, and Minimax Theory.

In the realm of Supervised Learning, participants will explore Linear Regression (covering low dimensional, ridge regression, lasso, and greedy regression), Nonparametric Regression (including kernel regression, local polynomials, additive models, and RKHS regression), Linear Classification (spanning linear models, logistic regression, SVM, and sparse logistic regression), Nonparametric Classification (covering methods like nearest neighbors (NN), naive Bayes, plug-in methods, and kernelized SVM), along with Conformal Prediction and Cross Validation techniques.

The Unsupervised Learning segment tackles Nonparametric Density Estimation, Clustering (k-means, mixtures, single-linkage, density clustering, and spectral clustering), Measures of Dependence, and Graphical Models (including correlation graphs, partial correlation graphs, and conditional independence graphs).

Moreover, the course covers an array of other essential topics such as Nonparametric Bayesian Inference, Bootstrap and subsampling techniques, Interactive Data Analysis, Robustness in statistical analysis, Active Learning approaches, Differential Privacy, Deep Learning, Distributed Learning, and techniques for Streaming data.

Designed for individuals keen on deepening their knowledge in artificial intelligence, this course falls under multiple categories including Artificial Intelligence Courses, Machine Learning Courses, Deep Learning Courses, Supervised Learning Courses, and Unsupervised Learning Courses. Offered independently, this program is a must for those aimed at broadening their expertise in statistical machine learning.