CSE 474: Introduction to Machine Learning

Humans have an uncanny ability to learn from their mistakes and adapt to new environments by relying on their past experience. Machine learning focuses on "How to write a computer program than can improve performance through experience?" Machine learning has a huge number of practical applications, more so in the present era of Big Data, where staggering volumes of diverse data in almost every facet of society, science, engineering, and commerce, are presenting opportunities for valuable discoveries. For example, machine learning is being used to understand financial markets, impact of climate change on society, protein-protein interactions, diseases, etc. Machine learning also has far ranging applications such as self-driving cars to never ending language learning systems.

This course will focus on understanding the mathematical and statistical foundations of machine learning. We will also cover the core set of techniques and algorithms needed to understand the practical applications of machine learning. The course will be an integrated view of machine learning, statistics (classical and Bayesian), data mining, and information theory. A basic understanding of probability, statistics, algorithms, and linear algebra is expected. Familiarity with Python is required for homework assignments and for understanding in-class demonstrations.

Following topics will be covered:
Concept learning: Hypothesis Space, Version Space
Computational learning theory: VC Dimension, PAC Learning
Supervised Learning: Instance based learning, margin-based classifiers, linear classification models, neural networks
Kernel Methods: Support Vector Machines
Unsupervised Learning: Clustering, Mixture of Gaussians