CS 550
Machine Learning

Description: Overview of Machine Learning. Concept Learning. Version Spaces. Inductive Bias. Induction of Decision Trees, overfitting, pruning. Evaluating Hypotheses. Bayesian Learning, Bayes Optimal Classifier, Naive Bayes Classifier, Bayesian Networks. Computational Learning Theory. Instance-Based Learning, k-Nearest Neighbor Learning, Locally Weighted Linear Regression. Genetic Algorithms, Genetic Programming. Learning Sets of Rules. Analytical Learning. Reinforcement Learning. Feature Projections Based Approaches. Credit units: 3.

Semester: Spring 2005
Schedule: Monday 11:40 - 12:30; Thursday 13:40 - 15:30
Office Hours: Wednesday 15:40 - 17:40
Classroom: EA-502
Instructor: H. Altay Güvenir

Main Text Book:
Tom M. Mitchell, Machine Learning, McGraw-Hill, 1997.

Recommended Text Book:
Pat Langley, Elements of Machine Learning, Morgan Kaufmann, 1996.

Recommended Journals:
Machine Learning, Journal of Machine Learning Research, Artificial Intelligence, Journal of Artificial Intelligence Research, IEEE Transactions on Pattern Analysis and Machine Intelligence. Knowledge-Based Systems

Weekly Schedule

WEEK  DAYS              TOPICS
   1           Feb.  3  Overview of Machine Learning, Concept Learning
   2  Feb.  7, Feb. 10  Concept Learning Version Spaces
   3  Feb. 14, Feb. 17  Decision Tree Learning, Evaluating Hypotheses
   4  Feb. 21, Feb. 24  Bayesian Learning, Naive Bayesian Learning
   5  Feb. 28, Mar.  3  Instance-Based Learning
   6  Mar.  7, Mar. 10  Genetic Algorithms
   7  Mar. 14, Mar. 17  Seminar
   8  Mar. 21, Mar. 24  Seminar
   9  Mar. 28, Mar. 31  Seminar
  10  Apr.  4, Apr.  7  Seminar
  11  Apr. 11, Apr. 14  Spring Break
  12  Apr. 18, Apr. 21  Seminar
  13  Apr. 25, Apr. 28  Workshop
  14  May   2, May   5  Workshop
  15  May   9, May  12  Workshop
  16  May  16           Workshop
Work Load

Seminar: Presentation: Each student will present a journal paper, or a small set of conference papers preferably written by the same group of authors on the same topic. Students are free to select the paper(s) they would like to present as long as they are in the scope of the course and approved by the instructor. Students will determine and submit, by email, the link to, or a hard copy of the paper(s) they will present to the instructor by March 1, 2005. The complete schedule of the presentations will be announced on March 3, 2005. Each student will have 20 mins. to present his/her paper. We will have a 5 min. discussion period after each presentation. Those who will present paper(s) that are available only on hard copies are responsible for providing the other classmates with copies of their paper(s) in a week advance.

Students are referred to the paper "How to Present a Paper in Theoretical Computer Science: A Speaker's Guide for Students" for a successful presentation.

Seminar: Participation: Other than his/her presentation, each student is expected to read the papers to be presented by others before the presentation. After each presentation, we will have a discussion session. Each student will be graded in his/her participation to the seminar.

Workshop: We will organize a workshop during the last weeks of the semester. Each student will conduct an experiment, testing new ideas preferably on the area of their presentation topic. Then each student will prepare a short paper, reporting his/her experiment(s) along with the interpretation of the results and pointers for further research. The papers must be in the range of 8-10 pages long, in 11 point Times New Roman font. The paper should have the quality of, at least, a national symposium paper. PDF versions of the papers will be submitted to the instructor by April 7, 2005. Each paper will be reviewed by two randomly selected classmates. The papers will be emailed for reviewing on April 8, 2005. Each reviewer will put his/her comments and suggestions on the paper and return them back to the instructor by April 18, 2005. The papers with peer reviews will then be returned to the authors on April 22, 2005. The workshop will start on April 25 at 14:40 and finish on May 16, 2005. Each student will have 15 min. to present his/her paper. The authors will revise, if necessary, their papers in the light of the reviews and discussions during the presentations, and submit the final copies by May 20, 2005 to the instructor, in PDF format.

Students are referred to the paper "Howto give a good research talk" for a successful presentation of your own work.

Grading Policy:

  Seminar:  40% (Presentation: 30% + Participation: 10%)
  Workshop: 60% (Presentation: 10% + Review: 10% + Paper: 40% )

On-line information:

Click here to send an email to whole class