ECE 66200 Pattern Recognition and Decision Making

Course Description: ECE 66200 Pattern Recognition and Decision Making

Textbook: Introduction to Statistical Pattern Recognition, K. Fukunaga

Notes: course notes

Homeworks:

• classifier equipped with different metrics
• metrics and kernel functions

Topics:

• Fisher kernel
• Reproducing kernel Hilbert spaces in Machine Learning

Blogs:

• distances / metrics
• Why is Euclidean distance not a good metric in high dimensions, some explanation
• statistical distance: f-divergence and KL, reverse KL, Pearson, reverse Pearson, Jensen-Shannon divergences, Hellinger, total variation distances
• Bregman divergence, Bregman distance, paper1, paper2
• concentration inequalities: Chernoff bound, Hoeffding’s, Pinsker’s inequality, Azuma’s and McDiarmid’s inequality
• reproducing kernel Hilbert spaces and transfer learning, domain adaption
• notes of Optimal Transport, arXiv summary paper
• deep metric learning: a (long) survey, a summary

References:

• The Elements of Statistical Learning, 2nd Edition
• Stochastic Processes: Theory for application, by R.G. Gallager, pdf
• R.G. Gallager’s lecture notes Chap3: detection, decisions, and hypothesis testing
• High-Dimensional Data Analysis with Low-Dimensional Models: Principles, Computation, and Applications