CSc 219 Machine Learning

Syllabus

Schedule

Project

Resources

Course Description

This course will introduce you to the major paradigms and methods of machine learning. Topics include: concept learning, decision tree learning, Bayesian learning, instance-based learning and case-based reasoning, genetic algorithms and genetic programming, learning sets of rules and inductive logic programming, analytical learning and explanation-based learning,  connectionist learning, ensemble learning, reinforcement learning and support vector machines. In the course, we will take a closer look at the characteristics, specific algorithms and tools, applicability and theoretical underpinnings of those learning methods (please refer to CSUS 2004 - 2006 Catalog, Page 232). As an integral part of the course, you will be required to complete a term project that include the following steps: (1) identify a real world problem, (2) select appropriate representation for both the training data and the knowledge to be learned, (3) collect data and prepare domain theory, (4) use some machine learning tool to carry out the learning process, and (5) analyze and evaluate the learned knowledge for the problem. Please refer to the Project page for detailed requirements.   3 units.

Prerequisites

The Catalog listed prerequisite is fully classified graduate status. You need to have maturity in problem solving (algorithm understanding and analysis), good programming skills, and  basic concepts of knowledge representations, search and performance analysis.

Classroom and Meeting Time:     DH 110,   MW: 5:30 - 6:45pm.

Office Hours:     MW 4:00 - 5:00pm,   or by appointment.

   Required:         Tom Mitchell, Machine Learning. McGraw-Hill, 1997.

   References:     

•   R. Sutton and A. Barto, Reinforcement Learning: An Introduction. MIT Press, 2nd ed., 1999.

•  J. R. Quinlan, C4.5: Programs for Machine Learning. Morgan Kaufmann Publishers, 1993.

•  B. Scholkopf and A.J. Smola, Learning with Kernels. MIT Press, 2002.

    Lecture notes and additional materials can be found in Schedule page. Please refer to Resources for pointers to supplemental materials.

1.   [30%]   Project

2.   [10%]   Homework

3.   [28%]   Midterm exam (close book/close notes)

4.   [32%]   Final exam (close book/close notes)

Check your grades here.

Policy on Letter Grade

The passing grade is B-.

Notes

1.   Any adjustment to this syllabus will be announced in class and posted on the web page. Class attendance is very important. Each student is responsible for any additional material to be discussed or distributed in class. There is a class mailing list (csc219) to be used for class-related discussion and posting. Subscription to the list is mandatory and you must do it during the first week of the semester.  A 0.5 percent will be deducted from the homework assignment category if your name is not on the list at the beginning of the second week of the semester.  Refer to the command list for the frequently used majordomo mailing list commands.

2.  Both exams will be close book/close notes exams. Prior to both exams, review guidelines will be posted and discussed. No make-up exam will be arranged unless there is a serious and compelling reason. Instructor must be notified prior to an exam.  

3.  Each homework is graded using 100 points. A late homework submission will result in 10 points reduction per day for no more than 5 school days after the due day. No credit will be given to any submission beyond the 5 school-day grace period. Late submission must be time-stamped at the Department Office (RVR 3018).

4.  The term project is a very important and integral part of the course. It is meant to be complementing and reinforcing classroom discussions and requires your earnest effort to accomplish. It is anticipated that you will spend a considerable amount of time outside classroom working on the project. Thus, it is very important that you start the project as early as possible. You are strongly encouraged to communicate with the instructor during the entire project development process.

5.  Cheating or plagiarism is a violation of a fundamental principle of academic honesty and integrity and will not be tolerated in the University. Under the provisions of the California Code of Regulations, cheating or plagiarism is cause for disciplinary action, including expulsion (Please refer to the University policy on plagiarism and academic dishonesty procedures). Your answers in the exams and assignments, and your report for the project must be your own. Since all parties involved will be subject to disciplinary action, you should be careful in guarding both printed and on-line versions of your work. Please read the Department Policy on Academic Integrity first, and then sign and return the Agreement sheet back to me.