Summary

This course introduces theoretical analysis of prediction methods, such as kernel methods and boosting algorithms, and examine questions about the performance guarantees we can provide for learning algorithms and the inherent difficulty of learning problems.

The topics to be covered are:

• Concentration inequalities;
• VC dimension, Sauer's lemma;
• Covering numbers and Rademacher averages;
• Generalization bounds for neural networks, kernel methods and voting classifiers

Lectures

Time: Monday/Wednesday 6:00 pm -7:15 pm; Location: Joshi 193

Instructor

Shaojun Wang
387, Joshi Center
shaojun.wang(at)wright.edu
(937) 775-5140
Office hours: Tuesday/Thursday 5:30PM-7:00PM

Textbooks

Vladimir N. Vapnik
Statistical Learning Theory
Wiley-Interscience, 1998.

Aad van der Vaart and Jon Wellner
Weak Convergence and Empirical Processes
Springer, 1996

Luc Devroye, Laszlo Gyorfi and Gabor Lugosi
A Probabilistic Theory of Pattern Recognition
Springer, 1996

Fan Chung and Linyuan Lu
Concentration Inequalities and Martingale Inequalities - A Survey
Internet Math., (3):79--127, 2007

Course Grades and Workload