Steam-powered Turing Machine University of Washington Department of Computer Science & Engineering
 CSE446 Course Overview
  CSE Home   About Us    Search    Contact Info 

Administrivia
 Home
 Using course email
 Email archive
 Policies
Content
 Overview
 Resources
 Lecture slides
Assignments
 Reading
 Homework
   

Tentative Order for Material to be Covered

  • Defining machine learning, applications, supervised, unsupervised, semisupervised and reinforcement learning, examples (polynomials, conjunctive concepts).
  • Key perspectives: function approximation, overfitting, bias, variance.
  • Random variables, probabilities, Bayes rule, MLE, MAP, conditional independence.
  • Multinomial naive Bayes, logistic regression, gradient ascent.
  • Evaluating learning systems
  • Generative and discriminative models, bias-variance decomposition, overfitting, regularization,
  • Feature selection, kernel functions, practical issues.
  • Probabilistic graphical models, inference (variable elimination, Gibbs sampling), learning Baysian networks.
  • Expectation maximization.
  • Hidden Markov models.
  • Back propagation, artificial neural networks, deep belief networks.
  • Support vector machines.
  • Semi-supervised learning, co-training.
  • Learning ensembles (bagging, stacking, boosting)
  • Learning theory
  • Clustering and dimensionality reduction.


CSE logo Department of Computer Science & Engineering
University of Washington
Box 352350
Seattle, WA  98195-2350
(206) 543-1695 voice, (206) 543-2969 FAX