Course Information

Course Coverage

This course will present an introduction to the field of machine learning and the related field of data mining. The course will cover material from both textbooks focusing primarily on Mitchell's book. The textbooks will be supplemented with material from research papers on recent developments in machine learning. The course will focus largely on classification learning, though there will also be some coverage of other learning techniques including analytical learning, unsupervised learning and reinforcement learning. The course will include several coding projects in which students will implement learning algorithms. Each student will also present a recent research paper during class and will undertake a significant course project.

Examinations, Assignments and Grades

Grades are assigned on a percentage basis, and then an adjustment is applied based on a minimum effort requirement (see below). The grade percentage cutoffs are as follows:

• A-minus cutoff is 90%
• B-minus cutoff is 80%
• C-minus cutoff is 70%
• D cutoff is 60%
• F is below 60%

These percentages may be lowered but will not be raised.

Minimum Effort Requirement: Students must turn in a minimal credible effort for EVERY assignment or their grade will be reduced one full letter grade (an A would become a B, an A- a B-, a B+ a C+, etc.). A turned in assignment achieving at least 40% of the possible points (before late assignment penalties) will be considered a minimal credible effort (though this percentage may be revised downwards by the instructor as warranted). For example, if a program has a maximum possible 30 points, then a turned-in assignment achieving at least 12 points before late penalties would be considered a minimal credible effort.

Final Term Project: each student will undertake to implement a major term project chosen in consultation with the instructor. Term projects may under special circumstances be done as team projects (but no more than two members to any team). Team projects should cover very ambitious implementation ideas that would ordinarily be well beyond the scope of a single student. A list of sample ideas will be posted as class goes on. Students will be expected to submit an initial (one page) outline for their project by February 17th. Further milestones will be annouced as class goes on. Each student will be expected to write up a short (no less than 6 and no more than 10 page) description of their implementation including the basic algorithm, experimental method, description of code. Each student will also make a short presentation of their implementation to the instructor during finals week.

Paper Presentation: each student will be expected to present a recent research paper chosen from a list assembled by the instructor (students may also suggest papers -- these suggestions must be improved by the instructor). Presentations will be 40 minutes in length and will begin starting in week 12 during class periods. Choice of paper to present will be given to students volunteering for the earliest presentation times. Students will also be expected to serve as a commentator for two other presentations (presenting five minutes of comments/questions/rebuttals on a paper). The final exam will cover the papers presented during class by the students.

Policies

Copies of the overheads used in class will be made available on the class web page.

You are responsible for what goes on in class, including lecture material, handouts, and turning in assignments. If you are unable to attend class it is your responsibility to obtain copies of class notes and any materials distributed in class. You may turn in copies of assignments early or have other members of the class turn in an assignment for you.

No exam will be given early. Exams can be made up only in the case of emergencies such as severe illness or death in the immediate family. You must contact me 24 hours in advance in order to arrange a makeup.

All assignments will be collected at the beginning of class on the due date. Late assignments will be penalized 20% of the grade for each working day the assignment is late.

Programming assignments must be your own work. You may discuss general ideas with other students, but should not discuss actual code with others. If you are having problems with an assignment, please come and see me or send me email.

As instructor I shall make every attempt to treat all students equally, without regard to race, religion, color, sex, handicap, age, veteran status, or sexual orientation. I encourage you to talk to me about your concerns of equal opportunity in the classroom. To inquire further about the University's policy on equal opportunity, contact the Office of Equal Opportunity (6827), 269-273 DAdB.

If you have any disability (either permanent or temporary) that might affect your ability to perform in this class, please inform me at the start of the semester. I may adapt methods, materials, or testing so that you can participate equitably. To learn about the services that UMD provides to students with disabilities, contact the Access Center (8727), 138 Kirby Plaza, or the Office of Equal Opportunity (8217), 269-273 DAdB.