Computing Science 466/551: Introduction to Machine Learning
Sept-Dec 2009 MEC 3-01 -- T Th 12:30-1:50pm
http://www.cs.ualberta.ca/~greiner/C-466/
Intructor: R Greiner [x2-5461, greiner@cs.ualberta.ca, Ath 359]
Barnabas Poczos [x8-1435, poczos@cs.ualberta.ca, Ath 405]
TAs: Shahab Jabbari Arfaee Gabor Gabor Balazs
{jabbaria | gbalazs}@cs.ualberta.ca
TEXTBOOKS:
Required: Hastie/Tibshirani/Friedman: The Elements of Statistical Learning:
Data Mining, Inference, and Prediction (2nd ed); Springer, 2009.
Rec'd: Pattern Recognition and Machine Learning, C.M. Bishop, Springer, 2006.
Alpaydin: Introduction to Machine Learning; MIT Press, 2004.
Duda/Hart/Stork: Pattern Classification; Wiley, 2001.
Mitchell: Machine Learning; McGraw Hill, 1997.
+ various URLs, handouts, ...
?? Reinforcement Learning, Sutton&Barto, online.
PURPOSE:
This course provides a (near)graduate-level introduction to the field of
Machine Learning, with an emphasis on the design on agents that can learn
about their environment, to help them improve their performance on a
range of tasks. We will cover both
(1) PRACTICAL ASPECTS, including algorithms for learning linear
classifiers, decision trees, SVMs, neural networks and belief networks;
and general models such as reinforcement learning; and
(2) THEORETICAL CONCEPTS, including relevant ideas from statistics,
inductive bias, Bayesian learning and the PAC learning framework.
The assignments will include both hands-on experiments with various learning
algorithms (linear units, neural networks, decision trees, reinforcement
learnings) -- many in C/C++, JAVA or Matlab -- as well as theoretical analysis.
Tentative OUTLINE:
Linear Classifiers/Regressors
Artificial Neural Networks
Decision Trees
Support Vector Machine; Kernel Methods
Evaluating Hypotheses (cross-validation)
Bias/Variance, Regularization, Model selection, Feature selection...
Computational Learning Theory
Bayesian Decision Theory
Ensemble Methods (Boosting, ...)
Unsupervised learning
Learning Graphical Models: Belief Nets, Markov Random Fields, ...
Reinforcement Learning
+ other topics, possibly including
Nearest neighbor; Learning Sets of Rules; ...
+ applications, possibly including
Adaptive Software, Computational Biology, ...
EVALUATION:
Assignments (tentatively: 4 -- both dry-lab + coding) (70% total)
Project (30%) including presentations,
+ meetings + write-up
Late Policy: Assignments are due at the beginning of class.
You will be "forgiven" a total of 4 late days for the assignments;
at most 2 for any single assignment.
OFFICE HOURS:
Immediately after class (2:00-2:30pm, T/Th) or by arrangement.