CS 269, Spring 2013: Topics in Machine Learning
Instructor: Zhuowen Tu
Time: 4:00PM~5:50PM, Tuesday and Thursday
Location: Bolter Hall 5273

webpage: http://www.loni.ucla.edu/~ztu/courses/2013_CS_spring/cs269_spring_2013.html

We will go through some of the topics in machine learning covering (1) supervied learning: ensemble-based discriminative learning (bagging, boosting, and random forest), (2) support vector machine (SVM) families (svm, structural svm, latent svm), (3) semi-supervised learning (transductive learning, graph Laplacian), (4) unsupervised manifold/metric learning, (5) sparsity and low-rank, (6) hierarchical models (deep learning, active basis, and biology driven models).

Grading policy: Two small projects with a medium-size project.

Covering topics:
Week 1, Tuesday (April 2):  Introduction of the course Introduction
Week 1, Thursday (April 4): Background introduction to discriminiative classifiers.
Classifier basics, VC dimension
Reading Materials
Machine learning review articles (P. Domingos)
Kernel Tricks (M. Jordan)
Classifier empirical studies (R. Caruana and A. Niculesu-Mizil, ICML 2006)
Classifers in high-dimensional data (Caruana et al., ICML 2008)

Week 2, Tuesday (April 9): Basics about SVM classifier Support Vector Machines

Week 2, Thursday (April 11): Supervised learning
Empirical study of the performance of  popular classifiers
Ensemble classifiers: bagging, boosting, random forests
Reading Materials
Pattern Classification (Chapter 4) (Duda)
Semi-supervised Learning Survey (X. Zhu)

Project 1
Due date: April 25

Week 3, Tuesday (April 16) Structrual prediction Label propagation, transductive learning, Multiple instance learning
Week 3, Thursday (April 18) Structrual prediction  PAC  theory, Active learning Slides
Week 4, Tuesday (April 13) Semi-supervised Learning

Reading Materials
Multiple instance learning

Week 4, Thursday (April 25) Semi-supervised Learning
Stacking, Cascade models
Week 5, Tuesday (April 30)

Week 5, Thursday (May 2) Unsupervised Learning Harmonic functions, spectrum clustering, Diffusion maps Slides
Project 2
Due date: May 23

Week 6, Tuesday (May 7) Unsupervised Learning Manifold learning, metric learning

Week 6, Thursday (May 9) Sparse coding Sparse coding, Low-rank
Reading Materials
Diffusion Maps (Coifman)
LLE (Roweis and Saul)
ISOMAP (Tenenbaum et al.)
Normalized Cuts (Shi et al.)
Label Propagation (Zhu et al.)
Laplacian Eigenmaps (Belkin and Niyogi)

Week 7, Tuesday (May 14) Sparse coding

Week 7, Thursday (May 16)
Dictionary learning
Week 8, Tuesday (May 21) Deep learning
RBM and deep belief network
Week 8, Thursday (May 23) Hierarchical models

Week 9, Tuesday (May 28) Big data

Week 9, Thursday (May 30)