CSC2515 Fall 2005 - Lectures

Tentative Lecture Schedule

• Sept 13 -- Machine Learning:
  Introduction to Machine Learning, Generalization and Capacity (notes [ps.gz] [pdf])

• Sept 20 --Classification 1:
  KNN, linear discriminants, decision trees (notes [ps.gz] [pdf])

• Sept 20 -- TUTORIAL (Prob/Stats/Linear Algebra Review/Questions)

• Sept 26 -- Classification 2:
  probabilistic classifiers: class-conditional Gaussians, naive Bayes, logistic regression, neural nets for classification (notes [ps.gz] [pdf])

• Oct 4: Assignment 1 (Classification) posted

• Oct 4 -- Regression 1:
  constant model, linear models, generalized additive models (e.g. RBFs), locally weighted regression, multilayer perceptrons/neural networks (notes [ps.gz] [pdf])

• Oct 11 -- Objective Functions and Optimization:
  error surfaces, weight space, gradient descent, stochastic gradient, conjugate gradients, second order methods, convexity, enforcing constraints (notes [ps.gz] [pdf])

• Oct 11 -- TUTORIAL (A1 Questions)
• Oct18: Assignment 1 due at the start of class

• Oct 18 -- Regression 2 and Supervised Mixtures:
  credit assignment problem, neural networks, radial basis networks, kolmogorov's theorem, backprop algorithm for efficiently computing gradients, mixtures of experts, piecewise models (notes [ps.gz] [pdf])

• Oct25: Assignment 2 (Regression) posted

• Oct 25 -- Unsupervised Learning 1: Trees & Clustering
  K-means, heirarchical clustering (alglomerative and divisive), maximum likelihood trees, optimal tree structure (notes [ps.gz] [pdf])

• Nov 1 -- Unsupervised Learning 2: Mixture models and the EM Algorithm:
  missing data, hidden variables, Jensen's inequality, lower bound on marginal likelihood, free energy interpretation, inference, (notes [ps.gz] [pdf])

• Nov 1 -- TUTORIAL (A2 Questions)

• Nov 8: Assignment 2 due

• November 8 -- Unsupervised Learning 3:
  Continuous latent variable models, Factor Analysis, (Probabilistic) PCA, Mixtures of Factor Analyzers, Independent Components Analysis (notes [ps.gz] [pdf])

• Nov 15: Assignment 3 posted

• Nov 15 -- Time Series Models
  autoregressive/Markov models, aggregate Markove models, hidden Markov models, profile HMMs (notes [ps.gz] [pdf])

• Nov 22 -- Capacity Control:
  generalization and overfitting, No free lunch theorems, high dimensional issues.
  capacity control methods: weight decay, early stopping, cross validation, model averaging, intro to Bayesianism (notes [ps.gz] [pdf])

• Nov 22 -- TUTORIAL (A3 Questions)

• Nov 29: Assignment 3 due at the start of class

• Nov 29 -- Meta-Learning Methods:
  stacking, bagging, boosting (notes [ps.gz] [pdf])

• Dec 2 (UNUSUAL TIME 10am, room UC163) -- Kernel methods:
  the kernel trick, support vector machines, kernel perceptrons, sparsity, capacity control, dual problems (notes [ps.gz] [pdf])

• Dec 6 -- NO CLASS (rescheduled to Dec2)

• December 19 -- projects due by email before 9am
  Send attachments or valid URL pointing to your report.
  POSTSCRIPT or PDF only. DO NOT SUBMIT WORD, HTML OR OTHER FORMAT FILES.

• December 19 -- all readings must be completed by 9am.
  Online Reading Submission

• Extra topics we may or may not have time for

CSC2515 - Machine Learning || www.cs.toronto.edu/~roweis/csc2515/