ENGN 2520 Pattern Recognition and Machine Learning

Spring 2012

Lecture: Tue/Thu 2:30pm-3:50pm

Location: Barus & Holley 159

Instructor: Pedro Felzenszwalb
Email: pff (at) brown.edu
Office: Barus & Holley 355
Office hours by appointment

Course description

This course covers fundamental topics in pattern recognition and machine learning. We will consider applications in computer vision, signal processing, speech recognition and information retrieval. Topics include: decision theory, parametric and non-parametric learning, dimensionality reduction, graphical models, exact and approximate inference, semi-supervised learning, generalization bounds and support vector machines. Prerequisites: basic probability, linear algebra, calculus and some programming experience.

Textbook

C. Bishop, Pattern Recognition and Machine Learning, Springer

Grading

Grading will be based on regular homework assignments and two exams. Homework will involve both mathematical exercises and programming assignments in Matlab. Students may discuss and work on homework problems in groups. However, each student must write down the solutions independently. Each student should write on the problem set the set of people with whom s/he collaborated.

Homework

Homework 1
Due Friday February 10 by 4pm
Data for programming assignment

Homework 2
Due Wednesday February 22 by 4pm
Data for programming assignment

Homework 3
Due Friday March 9 by 4pm

Homework 4
Due Friday April 6 by 4pm
Data for programming assignment

Homework 5
Due Thursday May 3 by 4pm
Data for programming assignment

Final exam

Take-home
Due Thursday May 10 by 4pm

Lectures

Lecture	Date	Topic	Reference (book sections)
1	January 26	Introduction and overview
2	January 31	Linear models for regression, basis functions, least squares	1.1, 3.1
3	February 2	Linear models for regression, regularization, probabilistic perspective	3.1.1, 3.1.2, 3.1.4
4	February 7	Classifiers, decision theory	1.5, 1.5.1
5	February 9	Decision theory, loss functions	1.5.2
6	February 14	Naive bayes classifier, ML estimation	4.2.2, 4.2.3
7	February 16	Linear threshold classifier, Perceptron	4.1.7
8	February 23	Linear Support Vector Machines	7.1, 7.1.1
9	February 28	Generalization bounds, VC-dimension	7.1.5
10	March 1	Kernels	6, 6.1
11	March 6	Sequential data	13, 13.1
12	March 8	Hidden Markov Models	13.2
13	March 13	HMM computation (forward/backward)	Rabiner
14	March 15	HMM computation (viterbi)	Rabiner
15	March 20	Bayesian networks	8.1, 8.2
16	March 22	Bayesian networks
17	April 3	Markov Random Fields	8.3.1, 8.3.2, 8.3.4
18	April 5	MRFs for image analysis	8.3.3
19	April 10	Markov chains	11.2
20	April 12	Gibbs sampling	11.3
21	April 17	Clustering, k-means	9.1
22	April 19	Mixtures of gaussians	9.2
23	April 24	Expectation Maximization	9.4
24	April 26	Principal Component Analysis	12.1
25	May 1	Linear Discriminant Analysis & Random Projections