Wei Lin @ PKU

00102721: Advanced Theory of Statistics II

Course Description

This course continues and extends the study of statistical theory initiated in Advanced Theory of Statistics. After introducing modern mathematical tools including empirical processes, concentration inequalities, and random matrices, we will discuss general theory and methodology for M- and Z-estimators, semiparametric and nonparametric models, and high-dimensional models, from both asymptotic and nonasymptotic points of view. Working through this course will help students develop a comprehensive understanding of modern statistical theory and acquire the necessary knowledge and skills for conducting research in statistics and data science.

Syllabus

Lectures and Assignments

Week Date Topics References Assignments and Notes
1 9/9 Overview, statistical motivations Wainwright Chap. 1, Sec. 4.1
9/11 Stochastic convergence in metric spaces, classical empirical processes vdV Chap. 18, Sec. 19.1
2 9/16 Empirical processes via bracketing and uniform entropy vdV Sec. 19.2, Kosorok Secs. 8.2–8.4
3 9/23 VC classes, preservation results, random functions vdV Secs. 19.2, 19.4, Kosorok Sec. 8.2, Chap. 9 Homework 1
9/25 No class
4 9/30 Functional delta method, Chernoff and sub-Gaussian tail bounds vdV Chap. 20, Wainwright Sec. 2.1
5 10/7 No class
10/9 Sub-exponential tail bounds, martingale concentration inequalities Wainwright Secs. 2.1, 2.2
6 10/14 Concentration for Lipschitz functions and empirical processes Wainwright Secs. 2.3, 3.3, 3.4
7 10/21 Covariance matrices, matrix functions Wainwright Secs. 6.1–6.4
10/23 No class, made up on 10/30
8 10/28 Concentration for random matrices, mimimax lower bounds Wainwright Secs. 6.4, 15.1 Homework 2
10/30 Le Cam's two-point method Wainwright Sec. 15.2
9 11/4 Generalized Le Cam, Fano's method Wainwright Secs. 15.2, 15.3, Tsybakov Sec. 2.6
11/6
10 11/11
11 11/18
11/20
12 11/25
13 12/2
12/4
14 12/9 Midterm exam
15 12/16
12/18
16 12/23