Machine learning is concerned with the question of how to construct programs that automatically improve their performance through experience. This course introduces the principles underlying the design of existing supervised and unsupervised machine learning algorithms. Topics to be covered include the following:
1. Supervised learning models and methods: Decision trees, neural networks, nearest-neighbor algorithms, Bayesian learning, Hidden Markov Models, support vector machines
2. Unsupervised learning: Clustering
3. Reinforcement learning: Markov Decision Processes, online supervised learning, certainty equivalent learning, temporal difference learning, Q-learning
4. General techniques: Feature selection, cross-validation, maximum likelihood estimation, expectation-maximization, gradient descent, ensemble learning
5. Statistical learning theory: Generalization error bounds, the PAC learning framework, VC dimension
Introduction
offering time
Fall 22
Major
Computer Science
Faculty
Vincent Ng(V)
Category
Course code