CS/ECE/STAT-861: Theoretical Foundations of Machine Learning

University of Wisconsin-Madison, Fall 2023

Course Schedule

Date Topics Scribed lecture notes Recommended reading Announcements
Wed. 9/6 Class overview and logistics,
PAC Learning, ERM
Course overview
Scribed Lecture 1 & 2
MRT Chapter 2,
SB Chapters 2, 3
Homework 0 released.
Fri. 9/8 PAC Learning (cont'd),
The agnostic case
Scribed Lecture 1 & 2 MRT Chapter 2,
SB Chapters 4
Mon. 9/11 Rademacher complexity Scribed Lecture 3 MRT Chapter 3
Wed. 9/13 Rademacher complexity,
Growth function
Scribed Lecture 4 MRT Chapter 3,
SB Chapter 6
Fri. 9/15 VC dimension
Scribed Lecture 5 MRT Chapter 3,
SB Chapter 6
Homework 0 due.
Mon. 9/18 PAC bound in a finite VC class,
Proof of Sauer's lemma
Scribed Lecture 6 MRT Chapter 3,
SB Chapter 6
Homework 0 solutions posted.
Homework 1 partially released.
Wed. 9/20 Lower bounds for point estimation,
Average risk optimality
Scribed Lecture 7 Lectures 7, 8, 9, 10 from
Lester Mackey's class
Fri. 9/22 Minimax optimality for point estimation
and beyond
Scribed Lecture 8 JD Chapter 2
Mon. 9/25 From estimation to hypothesis testing,
Neyman-Pearson test, Le Cam's method
Scribed Lecture 9 JD Chapter 7 Homework 1 updated.
Wed. 9/27 Le Cam's method examples Scribed Lecture 10 JD Chapter 7
Fri. 9/29 Review of Information theory Scribed Lecture 11 Cover & Thomas Chapter 2
Mon. 10/02 Fano's inequality, Fano's method
Constructing alternatives via tight packings
Scribed Lecture 12 JD Chapter 7
Wed. 10/04 Varshamov-Gilbert lemma,
Nonparametric regression
Scribed Lecture 13 AT Chapter 1.5, 2.5
Fri. 10/06 Nonparametric regression cont'd, Nadaraya-Watson
estimator, Nonparametric density estimation
Scribed Lecture 14 AT Chapter 1.5, 2.5, 1.2 Homework 1 due.
Mon. 10/09 Density estimation cont'd, Kernel density estimation,
Minimax lower bounds for prediction problems
Scribed Lecture 15 AT Chapter 1.2 Homework 1 solutions posted,
Homework 2 partially released.
Wed. 10/11 Classification in a VC class revisited: lower bounds
using Fano's method, Stochastic bandits introduction
Scribed Lecture 16 MRT Chapter 2,
LS Chapter 1, 2, 4
Fri. 10/13 The optimism in the face of uncertainty principle
The Upper Confidence Bound (UCB) algorithm
Scribed Lecture 17 LS Chapter 7
Mon. 10/16 Upper bounds for UCB,
Lower bounds for K-armed bandits
Scribed Lecture 18 LS Chapter 15, 16 Homework 2 updated.
Wed. 10/18 Lower bounds for K-armed bandits (cont'd),
Structured bandits
Scribed Lecture 19 LS Chapter 15, 16,
LS Chapter 19
Fri. 10/20 Structured bandits (cont'd),
Martingales review
Scribed Lecture 20 LS Chapters 19, 20,
Filippi et al, 2010.
Project proposals due.
Mon. 10/23 Martingale concentration and structured banditsScribed Lecture 21 LS Chapters 19, 20
Wed. 10/25 Online learning, The experts problem
The Hedge algorithm
Scribed Lecture 22 FO Chapter 7
Fri. 10/27 The experts problem (cont'd)
Adversarial bandits, EXP3
Scribed Lecture 23 FO Chapter 10,
LS Chapter 11
Homework 2 due.
Mon. 10/30 Adversarial bandits (cont'd)
Lower bounds for adversarial bandits, experts problem
Scribed Lecture 24 LS Chapter 11 Homework 3 partially released.
Wed. 11/01 Online convex optimization, Follow the
(regularized) leader, Failure cases for FTL
Scribed Lecture 25 FO Chapter 7, Haipeng
Luo's lecture notes
Homework 2 solutions posted
Fri. 11/03 Convexity review, FTRL with
strongly convex regularizers, Applications
Scribed Lecture 26 FO Chapter 7 Homework 3 updated
Mon. 11/06 Online gradient descent,
Contextual bandits, EXP4
Scribed Lectures 27 & 28 Haipeng Luo's notes,
LS Chapter 18
Wed. 11/08 Contextual bandits (cont'd),
Exam review and wrap up
Scribed Lectures 27 & 28 LS Chapter 18 Homework 3 due.
End of class
Wed. 11/15 Take-home exam from 11/14-11/17
Fri. 11/17 Take-home exam from 11/14-11/17
Fri. 12/8
Final projects due.