CS494/594: Schedule/Readings/Notes

CS494/594: Machine Learning

Schedule/Readings/Notes
Fall 2007

Subject to change. Check back frequently for updates.
Last updated: November 27, 2007
(Unless otherwise noted, the readings are from the Mitchell textbook.)

Date Topics Assigned Readings

Thurs. 8/23 Course Introduction
Intro. to Learning
Lecture slides

Tues. 8/28 Designing learning systems

(No lecture slides, but see Mitchell's Ch. 1 slides for related material, especially slides 11-21.)
Ch. 1

Thurs. 8/30 (Wrapup) Issues in Machine Learning
Neural Networks
Introduction

(No lecture slides, but see Mitchell's Ch. 4 slides for related material.)
Ch. 4.1-4.3

"The Basic Ideas in Neural Networks", by Rumelhart et al., Communications of the ACM, 37(3): 87-92, 1994.

Tues. 9/04 Neural Networks (con't.)
Perceptrons Ch. 4.4

Thurs. 9/6 Neural Networks (con't.)
Multi-layer feedforward neural networks
Back Propagation Ch. 4.5-4.6
Notes on Neural Networks
Project 1 assigned , due Oct. 1

Tues. 9/11 Neural Networks (con't.)
Convergence issues
K-fold cross-validation
Design and performance issues
Case studies
Ch. 4.7-4.9
Analysis of Hidden Units in a Layered Network Trained to Classify Sonar Targets, by Gorman and Sejnowski, Neural Networks, Vol. 1, pgs. 75-89, 1988.

Thurs. 9/13 In-class design exercise #1 (in teams):
Using neural networks to learn the paddle ball task

Tues. 9/18 Bayesian Learning
Review of basic probability
Introduction
(No lecture slides, but see Mitchell's Ch. 6 slides for related material.)
(See parts of Ch. 5.3)
Ch. 6.1-6.2

Thurs. 9/20 Bayesian Learning (con't.)
Bayes optimal classifier
Gibbs algorithm
Naive Bayes Classifier Ch. 6.7-6.9

Tues. 9/25 Bayesian Learning (con't.)
Example: text classification
Boosting, using AdaBoost Ch. 6.10
Extra handouts: from Russell and Norvig's Artificial Intelligence, pgs. 718, 664-668.
Boosting Naive-Bayes Classifiers to Predict Outcomes for Hip Prostheses, by Navone et al, IJCNN, 1999.

Thurs. 9/27 Bayesian Learning (con't.)
Bayesian Belief Networks
EM Algorithm Ch. 6.11 - 6.12

Tues. 10/2 Overview of Project 2
Reinforcement Learning
Introduction
Project 2 assigned, due Oct. 22
Ch. 13.1 - 13.2

Thurs. 10/4 Tutorial on Player/Stage
(to be used in Project 3;
taught by Rasko Pjesivac) Player Stage Getting Started Guide

Tues. 10/9 Reinforcement Learning (con't.)
Q learning for deterministic domains Ch. 13.3

Thurs. 10/11 No class (fall break)

Tues. 10/16 Reinforcement Learning (con't.)
Q learning in non-deterministic domains
TD Learning Ch. 13.4 - 13.5

Thurs. 10/18 Reinforcement Learning (con't.)
Sarsa (lambda)
Eligibility traces
Case Studies Extra handout: excerpts from Alpaydin's Introduction to Machine Learning, MIT Press, 2004, pgs. 383-397.

Improving Elevator Performance Using Reinforcement Learning, In Touretzky, et al (eds.), Advances in Neural Information Processing Systems: Proc. of the 1995 Conference, MIT Press, pgs. 1017-1023, 1996.
Extra handout: excerpts from Sutton and Barto's Reinforcement Learning: An Introduction, MIT Press, 1998, pgs. 261-267, 274-279.

Tues. 10/23 Genetic Algorithms
Project 3 assigned, Part A due Oct. 26; Full project due Nov. 12
Ch. 9.1

Thurs. 10/25 Genetic Algorithms Ch. 9.2-9.3

Tues. 10/30 In class design exercise #2 (in teams):
Learning to play blackjack (21) using
reinforcement learning and genetic algorithms

Thurs. 11/1 No class; instructor on travel

Tues. 11/6 Genetic Programming Ch. 9.5
Extra handouts: excerpts from John Koza's Genetic Programming, parts of Ch. 6 (pgs. 79-94) and Ch. 7 (pgs. 147-187)

Thurs. 11/8 Genetic Programming Koza handouts from 11/6 class

Tues. 11/13 Genetic Programming
Review "quiz" contest
Koza handouts from 11/6 class
Project 4 assigned, due Dec. 3

Thurs. 11/15 Evolutionary programming example: GOLEM project for evolving robots
Instance-Based Learning
   K-Nearest Neighbor
   Weighted K-Nearest Neighbor
Automatic design and manufacture of robotic lifeforms, Nature, vol. 406, August 31, 2000, pg. 974-978.
Ch. 8.1-8.2

Tues. 11/20 Instance-Based Learning (con't.)
   Locally weighted regression
   Radial basis functions Ch. 8.3-8.4
Project 5 (poster) assigned; to be presented in Poster Session during final exam period (Dec. 6)

Thurs. 11/22 No class (Happy Thanksgiving!)

Tues. 11/27 Theoretical issues in machine learning
   Inductive learning hypothesis
   Version spaces
   Inductive bias Ch. 2.1-2.3, 2.5.1, 2.6.3, 2.7-2.8

Thurs. 11/29 Theoretical issues in machine learning (con't.)
   PAC learning model Ch. 7

Tues. 12/4 Theoretical issues in machine learning (con't.) Ch. 7

Thurs. 12/6
10:15 AM - 12:15 PM Project 5 Poster Session (during scheduled final exam time period), in Claxton Commons area

Date	Topics	Assigned Readings
Thurs. 8/23	Course Introduction Intro. to Learning Lecture slides
Tues. 8/28	Designing learning systems (No lecture slides, but see Mitchell's Ch. 1 slides for related material, especially slides 11-21.)	Ch. 1
Thurs. 8/30	(Wrapup) Issues in Machine Learning Neural Networks Introduction (No lecture slides, but see Mitchell's Ch. 4 slides for related material.)	Ch. 4.1-4.3 "The Basic Ideas in Neural Networks", by Rumelhart et al., Communications of the ACM, 37(3): 87-92, 1994.
Tues. 9/04	Neural Networks (con't.) Perceptrons	Ch. 4.4
Thurs. 9/6	Neural Networks (con't.) Multi-layer feedforward neural networks Back Propagation	Ch. 4.5-4.6 Notes on Neural Networks Project 1 assigned , due Oct. 1
Tues. 9/11	Neural Networks (con't.) Convergence issues K-fold cross-validation Design and performance issues Case studies	Ch. 4.7-4.9 Analysis of Hidden Units in a Layered Network Trained to Classify Sonar Targets, by Gorman and Sejnowski, Neural Networks, Vol. 1, pgs. 75-89, 1988.
Thurs. 9/13	In-class design exercise #1 (in teams): Using neural networks to learn the paddle ball task
Tues. 9/18	Bayesian Learning Review of basic probability Introduction (No lecture slides, but see Mitchell's Ch. 6 slides for related material.)	(See parts of Ch. 5.3) Ch. 6.1-6.2
Thurs. 9/20	Bayesian Learning (con't.) Bayes optimal classifier Gibbs algorithm Naive Bayes Classifier	Ch. 6.7-6.9
Tues. 9/25	Bayesian Learning (con't.) Example: text classification Boosting, using AdaBoost	Ch. 6.10 Extra handouts: from Russell and Norvig's Artificial Intelligence, pgs. 718, 664-668. Boosting Naive-Bayes Classifiers to Predict Outcomes for Hip Prostheses, by Navone et al, IJCNN, 1999.
Thurs. 9/27	Bayesian Learning (con't.) Bayesian Belief Networks EM Algorithm	Ch. 6.11 - 6.12
Tues. 10/2	Overview of Project 2 Reinforcement Learning Introduction	Project 2 assigned, due Oct. 22 Ch. 13.1 - 13.2
Thurs. 10/4	Tutorial on Player/Stage (to be used in Project 3; taught by Rasko Pjesivac)	Player Stage Getting Started Guide
Tues. 10/9	Reinforcement Learning (con't.) Q learning for deterministic domains	Ch. 13.3
Thurs. 10/11	No class (fall break)
Tues. 10/16	Reinforcement Learning (con't.) Q learning in non-deterministic domains TD Learning	Ch. 13.4 - 13.5
Thurs. 10/18	Reinforcement Learning (con't.) Sarsa (lambda) Eligibility traces Case Studies	Extra handout: excerpts from Alpaydin's Introduction to Machine Learning, MIT Press, 2004, pgs. 383-397. Improving Elevator Performance Using Reinforcement Learning, In Touretzky, et al (eds.), Advances in Neural Information Processing Systems: Proc. of the 1995 Conference, MIT Press, pgs. 1017-1023, 1996. Extra handout: excerpts from Sutton and Barto's Reinforcement Learning: An Introduction, MIT Press, 1998, pgs. 261-267, 274-279.
Tues. 10/23	Genetic Algorithms	Project 3 assigned, Part A due Oct. 26; Full project due Nov. 12 Ch. 9.1
Thurs. 10/25	Genetic Algorithms	Ch. 9.2-9.3
Tues. 10/30	In class design exercise #2 (in teams): Learning to play blackjack (21) using reinforcement learning and genetic algorithms
Thurs. 11/1	No class; instructor on travel
Tues. 11/6	Genetic Programming	Ch. 9.5 Extra handouts: excerpts from John Koza's Genetic Programming, parts of Ch. 6 (pgs. 79-94) and Ch. 7 (pgs. 147-187)
Thurs. 11/8	Genetic Programming	Koza handouts from 11/6 class
Tues. 11/13	Genetic Programming Review "quiz" contest	Koza handouts from 11/6 class Project 4 assigned, due Dec. 3
Thurs. 11/15	Evolutionary programming example: GOLEM project for evolving robots Instance-Based Learning K-Nearest Neighbor Weighted K-Nearest Neighbor	Automatic design and manufacture of robotic lifeforms, Nature, vol. 406, August 31, 2000, pg. 974-978. Ch. 8.1-8.2
Tues. 11/20	Instance-Based Learning (con't.) Locally weighted regression Radial basis functions	Ch. 8.3-8.4 Project 5 (poster) assigned; to be presented in Poster Session during final exam period (Dec. 6)
Thurs. 11/22	No class (Happy Thanksgiving!)
Tues. 11/27	Theoretical issues in machine learning Inductive learning hypothesis Version spaces Inductive bias	Ch. 2.1-2.3, 2.5.1, 2.6.3, 2.7-2.8
Thurs. 11/29	Theoretical issues in machine learning (con't.) PAC learning model	Ch. 7
Tues. 12/4	Theoretical issues in machine learning (con't.)	Ch. 7
Thurs. 12/6 10:15 AM - 12:15 PM	Project 5 Poster Session (during scheduled final exam time period), in Claxton Commons area

CS494/594: Machine Learning

Schedule/Readings/Notes Fall 2007

Schedule/Readings/Notes
Fall 2007