University of Chicago: Computer Science: CMSC 25000: Winter 2003: Course Description

CMSC 25000 - Artificial Intelligence
Winter 2003
Course Description

Description

Principles of artificial intelligence. Reasoning, Learning, and Systems. Knowledge representation and manipulation. Language, vision, and speech.

Goals

Understand techniques for search: heuristic and optimal.
Understand the interaction between knowledge representation and reasoning.
Analyze, evaluate, and apply machine learning techniques: nearest neighbor, identification trees, perceptrons and neural networks, and Hidden Markov Models.
Examine application of learning and search techniques to human skills: natural language understanding, speech processing, vision, and robotics.

Instructional Staff

	Instructor	Teaching Assistant
Name	Gina Levow	Dinoj Surendran
Email	levow@cs.uchicago.edu	dinoj@cs.uchicago.edu
Office	Ry 166	Eck 002a
Office Hours	Thursday 2:30-4:30 PM	Monday 3-4; Wednesday 4-5PM
Office Phone	x2-5680	x4-3615

Course Mechanics

A syllabus that summarizes what we will cover each week and gives a preview of the homework assignments and a reading list showing the reading for each week can be found on the class web page.

The instructors' office hours are for discussing material from class. Students wishing to discuss accommodations for unusual circumstances should also come see me, and should do so at the earliest possible time. Because many homework questions come up over and over, it is usually best to check with the TA before coming by during office hours with a homework question, but we will be happy to discuss the homework with you as well if office hours are the most convenient time to get your question resolved. Homework grades should always be discussed with the TA first, but your instructor will be happy to discuss them with you after you have done that if your concern is not resolved. We will hold scheduled office hours as noted above, but I will also be happy to meet with students that don't find that time convenient either after class or by appointment. Electronic mail is the best way to reach us to set up an appointment, and it is also a good way to get a quick answer to a simple question.

Course Materials

The class web site at http://classes.cs.uchicago.edu/classes/archive/2003/winter/cs25000/ contains the most recent copies of all material produced for this course. This course description (which contains a number of useful links), the syllabus, lecture notes and homework solutions can be found there.

The required text is Russell and Norvig's Artificial Intelligence: A Modern Approach, 2nd edition, available from the Seminary Co-op Bookstore.

Grading

Course grades will be assigned based on homework, an in-class midterm examination, and the final examination. Scores on each component will be combined to produce a single overall score for each student as follows:

Component	Percentage	Computation
Midterm and Final	60%	30%, and 30%
Homework	40%	5% each

Scores for each course requirement (homework assignment, exam ) will be assigned on a 100 point scale. Overall course grades will be assigned using a relative grading method in which the overall scores for all students will be used to produce a histogram and then breakpoints will be selected between the letter grades. The breakpoints are not predetermined, and will not be selected until after the histogram is prepared. Approximately one third of the students have earned some type of A and a small number of students have received grades below C.

Homework is due before the start of class. Credit for partial work will be given. Late assignments will be corrected, but will lose 10% for each 24 period after the due date. All material included in the homework is testable, so skipping an assignment is a bad idea.

In general, the TA will grade the homeworks under the supervision of the faculty. If you have a question about a homework grade, you should contact the TA first, but you may appeal to the faculty. Open-ended questions on the exams will be graded jointly by the faculty and TA.

Students may work together on the assignments, but all of the material which is turned in for grading must be produced individually. For example, students may form study groups and work out solutions together on a chalkboard, but it would not be permissible for one student to create a computer file containing the answers and then for other students to copy that file and submit it as their own work. The goal of this policy is to encourage the use of homework as a learning aid. Credit should be given for help on the homework by identifying one's collaborators on the first page of the submitted homework assignment.