Synopsis

Machine learning is an important and rapid growing branch of artificial intelligence. The aim of machine learning is to design algorithms that can extract information from environment automatically and improve their ability to perform the intended task. This course starts with a high level overview of general problems in machine learning, followed by a review of mathematical backgrounds and numerical optimization methods that are essential for machine learning algorithms, after that several important topics in machine learning will be covered. 

Lectures

• Class information [slides]
• Introduction [slides]
• Jupyter Notebook Tutorial [slides]
• Mathematical Backgrounds 
• Review of Linear Algebra (1) -- solving linear equation [slides]
• Review of Multivariate Calculus [slides]
• Review of Linear Algebra (2) -- eigenvalue problems [slides]
• Gradient descent and Newton method [slides]
• Unit 1: Linear Least Squares -- solving linear equations
• Regression and Linear Least Squares [slides]
• Robust learning and Reweighted Linear Least Squares [slides]
• Online learning and Recursive Linear Least Squares [slides]
• Model selection for LLSE [slides]
• Regularized LLSE [slides]
• LLSE for Classification [slides]
• LLSE for Ranking [slides]
• Multi-modal LLSE and k-means clustering [slides]
• Unit 2: Eigenvalue-based Methods 
• Total LLSE [slides]
• Principal Component Analysis [slides]
• Multi-dimesional Scaling and ISOMAP [slides]
• Spectral Clustering [slides]
• Unit 3: Classification Methods 
• Fisher Linear Discriminant [slides]
• Classification metrics [slides]
• Logistic regression [slides]
• Support vector machines
• Theory [slides]
• Algorithm [slides]
• Kernel SVM [slides]
• Unit 4: Neural Networks 
• Basic Neural Networks [slides]
• Deep Learning [slides]