285 -
Special Topics in Robotics and Control Systems

Statistical Learning I
 

 

             

This course provides an introduction to pattern recognition and statistical learning. Topics covered include: Bayesian decision theory; parameter estimation; maximum likelihood; the bias-variance trade-off; Bayesian parameter estimation; the predictive distribution; conjugate and non-informative priors; dimensionality and dimensionality reduction; principal component analysis; Fisher's linear discriminant analysis; density estimation: parametric vs kernel-based methods; mixture models; expectation-maximization; applications.

Lectures: TuTh, 3:30-5:00, HSS 1128A
Instructor: Nuno Vasconcelos
n u n o @ e c e . u c s d . e d u, EBU1-5603
office hours: Wed 3-4:30 PM
Text: Pattern Classification (2nd ed.)
R. Duda, P. Hart, and D. Stork, Wiley Interscience, 2000
Syllabus: [ps, pdf]
Homework: Problem set 1 [ps, pdf, data]
Problem set 2 [ps, pdf, data]
Problem set 3 [ps, pdf, data]
Problem set 4 [ps, pdf]
Problem set 5 [ps, pdf]
Problem set 6 [ps, pdf]
Cheetah Day: [ps, pdf]
Readings: Lecture 1: introduction (DHS, chapter 1)
Lecture 2: Bayesian decision theory (DHS, chapter 2)
Lecture 3: Bayesian decision theory (DHS, chapter 2)
Lecture 4: Gaussian classifier (DHS, chapter 2)
Lecture 5: Maximum-likelihood estimation (DHS, chapter 3)
Lecture 6: Bias and variance (DHS, chapter 9)
Lecture 7: Bias and variance (DHS, chapter 9)
Lecture 8: Bayesian parameter estimation (DHS, chapter 3)
Lecture 9: Bayesian parameter estimation (DHS, chapter 3)
Lecture 10: Conjugate and non-informative priors
Lecture 11: Dimensionality  (DHS, chapter 3) [slides]
Lecture 12: Dimensionality reduction (DHS, chapter 3)
Lecture 13: Nearest neighbor classifier (DHS, chapter 4)
Lecture 14: mid-term
Lecture 15: Kernel-based density estimates (DHS, chapter 4) [slides]
Lecture 16: Mixture models [slides]
Lecture 17: Expectation-maximization [slides]
Lecture 18: Expectation-maximization
Lecture 19: Applications [slides]