COMS 4773 Fall 2026

COMS 4773 is a graduate-level introduction to machine learning theory.

Course information

Time: Tue/Thu 1:10pm–2:25pm
Venue: 413 Kent Hall
Instructor: Daniel Hsu
Teaching assistants: Noah Bergam, Nikhil Mudumbi
Office hours:
- Mon 1:00pm–3:00pm (Noah, Mudd 4th Floor Data Science Suite, near the kitchenette)
- Tue 2:35pm–4:35pm (Daniel, Mudd 4th Floor Data Science Suite, room 426)
- Wed 12:00pm–2:00pm (Nikhil, CS Student Learning Center, 3rd floor Schermerhorn)
Links: Syllabus, Courseworks, EdStem, Gradescope

Check EdStem for important announcements.

09/08: Course overview; Using expert advice
- PS0, due 9/17 at 11:59pm
09/10: Using expert advice
- Reading: Avrim Blum’s survey article (Section 2)
09/15: Online allocation
- Reading: Freund and Schapire’s paper on Hedge (Sections 2 and 3)
09/17: Online allocation
- PS1, due 9/29 at 11:59pm
- Optional: Freund and Schapire’s paper on zero-sum games (Sections 5 and 6)
09/22: Online allocation, Online concept learning
- Reading: Avrim Blum’s survey article (beginning of Section 3 (OK to skip “agnostic learning” and “random noise” parts); Section 3.1)
09/24: Online concept learning
- Optional: Zuev’s paper on the number of LTFs (AI-generated English translation of Zuev’s paper)
09/29: Online concept learning
- Reading: Littlestone’s paper on mistake-bound learning (Sections 1–3)
10/01: Statistical learning
- Peer Evaluation assignments for PS1 out, due 10/8 (in lecture, or give to instructor/TA)
- PS2, due 10/15 at 11:59pm
- Optional: Littlestone’s paper on on-line to batch conversion; Berk and Freedman’s paper on statistical assumptions
10/06: Statistical learning
- Reading: Probability review handout
10/08:
10/13:
10/15:
10/20:
10/22:
10/27:
10/29:
11/03: Holiday (no lecture)
11/05: Midterm exam (in-class)
11/10:
11/12:
11/17:
11/19:
11/24:
11/26: Holiday (no lecture)
12/01:
12/03:
12/08:
12/10:
12/22: Final exam (at noon; registrar’s projection)