Semester: |
2025 Semester 2, also offered on 2025 Semester 1, 2024 Semester 2 and 2024 Semester 1 |
Time and place: |
The University of Melbourne, Monday 2.30-5pm, Melbourne Connect 5210 (except Monday October 13 10.30am-1pm, Melbourne Connect 3101) |
Organizer: | Jean Honorio |
Date | Topic (Tentative) | Notes |
Mon, Sep 15 |
Lecture 1: Markov's inequality, Chebyshev's inequality Lecture 2: Hoeffding's inequality, empirical risk minimization with a finite hypothesis class |
|
Mon, Sep 22 |
Lecture 3: Fano's inequality, empirical risk minimization with a finite hypothesis class (lecture continues) |
|
Mon, Sep 29 |
Lecture 4: probably approximately correct (PAC) Bayes, structured prediction (lecture continues) |
|
Mon, Oct 6 |
Lecture 5: McDiarmid's inequality, sub-Gaussian random variables Lecture 6: Rademacher complexity, linear prediction |
|
Mon, Oct 13 |
(lecture continues) Lecture 7: deterministic and stochastic optimization, convergence rates |
|
Mon, Oct 20 |
(lecture continues) Lecture 8: restricted strong convexity, compressed sensing |
|
Mon, Oct 27 |
(lecture continues) Lecture 9: primal-dual witness method, support recovery |
|
Mon, Nov 3 |
(lecture continues) Lecture 10: growth function, Vapnik-Chervonenkis (VC) dimension, Sauer-Shelah lemma, Massart lemma |
|
Mon, Nov 10 |
(lecture continues) Lecture 11: Le Cam's lemma |