CMSC848P: Machine Learning Theory

This course page is still under construction. I hope the current information is helpful as you plan and enroll in courses.

Key information

Time and location: Mon/Wed 3:30pm–4:45pm, CSI 1121
Instructor: Han Shao, hanshao@umd.edu
TA: TBD

Coursework and Grading (tentative)

4-5 homework assignments (30%), midterm exam (25%), final project (15%), final exam (25%).

Course description

This course is fully lecture based. It focuses on foundational tools in learning theory (e.g., generalization in the offline setting and regret bounds in the online setting) and explores active research directions. Machine learning theory asks questions like: what guarantees can we prove for practical learning methods, and can we design algorithms that achieve these guarantees? what can we say about the inherent ease or difficulty of different learning problems?

Prerequisites

Mathematical maturity and comfort with theorems/proofs are required. Familiarity with probability/statistics (e.g., concentration inequalities, union bound) and basic algorithms is expected. No programming is required. All homeworks and exams will consist of proof-based questions.

Homeworks

• You will be given a question set (usually) two weeks before the deadline.
• On the deadline date, either Monday or Wednesday, at the beginning of class, we will randomly pick one question and ask you to finish it in class.
• No access to any device and any notes during the in-class test.
• Your homework will be graded only based on the in-class submission.
• If you cannot attend the in-class test, please inform us at least 48 hours in advance.

Exams

Exams are considered "Major Scheduled Grading Events," a self-signed note may not be sufficient: For medical absences, you must furnish documentation from the health care professional who treated you, which must verify the timeframe that the student was unable to meet academic responsibilities. In addition, it must contain the name and phone number of the medical service provider to be used if verification is needed. No diagnostic information will ever be requested.

Reference material

• S. Shalev-Shwartz and S. Ben-David, Understanding Machine Learning: From Theory to Algorithms (UML)
• A. Blum, J. Hopcroft, and R. Kannan, Foundations of Data Science
• O. Bousquet, S. Boucheron, and G. Lugosi, Introduction to Statistical Learning Theory

Office hours

Han Shao: TBD, IRB 5132
TA: TBD

Schedule (tentative and subject to change)

Date	Topic
08/31/2026	Introduction
09/02/2026	PAC learning
09/07/2026	No class — Labor Day
09/09/2026	Sample complexity: upper bounds
09/14/2026	Sample complexity: lower bounds
09/16/2026	Agnostic learning
09/21/2026	Rademacher complexity
09/23/2026	PAC-Bayes
09/28/2026	Hardness of learning
09/30/2026	Regression
10/05/2026
10/07/2026
10/12/2026	No class — Fall Break
10/14/2026	Midterm exam
10/19/2026	Online learning: mistake bounds and Littlestone dimension
10/21/2026	Agnostic online learning; multiplicative weights and regret minimization
10/26/2026	Linear classes and the perceptron algorithm
10/28/2026	Boosting
11/02/2026
11/04/2026	Learning and games
11/09/2026	Distribution learning
11/11/2026	AI alignment
11/16/2026	Robustness
11/18/2026	One-inclusion graph
11/23/2026
11/25/2026	No class — Thanksgiving Recess
11/30/2026	Fairness
12/02/2026	Privacy
12/07/2026
12/09/2026
12/16/2026	Final exam, 4:00pm–6:00pm, based on the Fall 2026 Final Exam Schedule