Machine Learning (COMP-652)
Fall 2012

Lecture Schedule

Date Topic Materials
Sep.6 Introduction. Types of machine learning. Linear regression. Overfitting. Lecture 1 slides
Bishop, Sec. 1.1
NY Times Article on Statistics
If you need to catch up on the math, a brief probability review and linear algebra review from Stanford University
Sep.11 Overfitting, cross-validation and bias-variance error decomposition. Linear models with basis functions. Gradient descent Lecture 2 slides
Bishop, Sec. 3.1, 3.2
Sep.13 More on linear methods for regression. Analysis of least-squares as maximum-likelihood learning. L2 and L1 regularization. Bayesian learning Lecture 3 slides
Bishop, Sec. 3.1, 3.3
Sep.18 More on regularization. Classification. Generative vs. discriminative learning. Logistic regression Lecture 4 slides
Bishop, Sec. 4.2.
Sep.20 Logistic regression. Feed-foreward neural networks. Lecture 5 slides
Bishop, Sec. 4.3
Sep.25 Naive Bayes. Gaussian Discriminant Analysis Lecture 6 slides
Bishop, Sec. 5.1, 5.2, 5.3, 5.4, 5.5
Sep.27 Instance-based (non-parametric) learning Lecture 7 slides
Bishop, Sec. 2.5
Oct.2 Decision trees Lecture 8 slides
Bishop, Sec. 1.6, 14.4
Oct. 4 Support vector machines. Kernels Lecture 9 slides
Oct. 9 Kernelizing other algorithms. Wrap-up of SVMs Lecture 10 slides
Bishop, Sec. 4.1, 7.1.
See also the kernel machines web site
Oct.16 Ensemble methods. Bagging. Boosting Lecture 11 slides
Bishop, Sec. 6.1, 6.2, 7.1
See also the kernel machines web site
Oct. 18 Experimental analysis. Active learning Lecture 12 slides
Active learning tutorial by Burr Settles (2010) Mitchell, Chapter 5
See also ROC Graphs: Notes and Practical Considerations for Researchers by Tom Fawcett (2004).
Oct. 23 and 25 Computational learning theory Lecture 13 and 14 slides
Oct. 30 Clustering Lecture 15 slides
Bishop, Sec. 9.1
Nov. 1 Midterm exam. Covering lectures 1-13. You are allowed one double-sided "cheat sheet"
Midterm from 2011.
Midterm from 2009
Nov. 6 Expectation Maximization (EM) and mixture of Gaussians. Introduction to dimensionality reduction Lecture 16 slides
Bishop, Sec. 9.2, 9.3, 9.4
Nov. 8 PCA, Kernel PCA and other dimensionality reduction methods. Lecture 17 slides
Bishop, Sec. 12.1, 12.3, 12.4
Nov. 15 Time series data. Hidden Markov Models Lecture 18 slides
Bishop, Sec. 13.1, 13.2
See also HMM tutorial by Rabiner (1989)
Nov. 20 More on time series data. Kalman and particle filters. Bayes nets. Importance sampling. Gibbs sampling
Bishop, Sec. 8.1, 8.2, 13.3
See also Particle filter tutorial by Arulampalam et al (2002)
See also graphical models tutorial by Koller et al, 2007.
Nov. 13 Introduction to graphical models. Undirected models. Belief propagation
Bishop, Sec. 8.3, 8.4
See also graphical models tutorial by Koller et al (2007).
Nov. 22 MCMC methods in general. Learning in graphical models revisited

Nov. 27 Reinforcement learning
Nov. 29 More reinforcement learning

Dec. 4 Frontiers and current trends in machine learning