Computer Science 577

Seminar in AI: Optimization

Fall 2008

Instructor: Dan Gildea office hours Wed 1-2pm
Time: M/W 2:00-3:15, CSB 601
Motto: Meliora

Required text: Boyd and Vanderberghe, Convex Optimization

In addition to this text, we will cover additional numerical methods taken from Nocedal and Wright and Numerical Recipes. The second half of the course will focus on applications of optimization in the context of machine learning.

W 9/3 IntroBoyd ch 1
M 9/8 Convex SetsBoyd ch 2
W 9/10 Convex FunctionsBoyd ch 3
M 9/15 Convex Optimization ProblemsBoyd ch 4
W 9/17 Convex Optimization ProblemsBoyd ch 4
M 9/22 DualityBoyd ch 5Tag
W 9/24 Approximation and Statistical EstimationBoyd ch 6, 7Michael
M 9/29 Geometric ProblemsBoyd ch 8Chetan
W 10/1 Unconstrained MinimizationBoyd ch 9Satyaki
M 10/6 Fall break
W 10/8 Equality Constrained MinimizationBoyd ch 10
M 10/13Interior Point MethodsBoyd ch 11Chetan
W 10/15Interior Point MethodsBoyd ch 11Satyaki
M 10/20BFGSNocedal ch 8Chetan
W 10/22BFGS implementationNocedal ch 9, Daume
M 10/27Simulated AnnealingKirkpatrickLam
W 10/29Numerical Methods 1Press 10.0 - 10.3Yi
M 11/3 Numerical Methods 2Press 10.4 - 10.6Yi
W 11/5 Integer ProgrammingNemhauser ch II.4Yi
M 11/10Machine Translation: MERTOch 2003, CerTag
W 11/12Making large scale SVM learning practicalJoachimsMichael
M 11/17Structured Max-MarginTaskar 2003 Taskar 2004
W 11/19SMO for Structured Max-MarginTaskar
M 11/24Transductive SVMCollobert
W 11/26Thanksgiving break
M 12/1 Inducing features of random fieldsDella PietraTag
W 12/3 Exponentiated GradientBartlettSatyaki
M 12/8 Dirichlet Process IntroRanganathanLam
W 12/10Variational Inference for Dirichlet Process MixturesBleiLam
Projects due: 5pm Friday 12/19

Grading

  • Lectures and participation: 60%
  • Final project: 40%

gildea @ cs rochester edu
January 7, 2009