| Artificial Intelligence |
| Course |
CS480 Sections 001 and 002 |
| Prerequisites |
CS310 and CS330 |
| Instructor |
Sean Luke,
415 S&T II, 3-4169 |
| Instructor Office Hours |
Mondays 10:10-11 pm, Tuesdays 7:20-9:00 pm |
| TA |
Shen Shyang Ho(sho@gmu.edu)
|
| TA Office Hours |
Tuesdays 2-4 pm, Thursdays 3-5 pm, Room 435 |
| Section 001 |
Robinson B208, Tuesdays 4:30 - 7:10 PM |
| |
| Section 002 |
Robinson B118, Mostly Mondays 7:20 - 10:00 PM |
| |
About the Course
This course will begin by covering the basics of Lisp and the philosophy of Artificial Intelligence, plus discussion of simple systems, architectures, and platforms (robotics, etc.). From there we will discuss methods in learning (neural networks, decision trees, optimization, and time permitting, reinforcement learning). Then the course will turn to issues in problem solving and search, game design, and logic and representation.
This course will be very challenging but (I hope!) interesting and eye-opening. Artificial Intelligence is a broad interdisciplinary field with a strong tradition in exploratory programming. You are expected to know the material in CS310 and CS330 well, and be able to get up to speed rapidly doing software development with strange new programming languages. Learning Lisp is a nontrivial endeavor. You should also be prepared to discuss and think about philosophical issues and be able to draw ideas from areas outside of computer science.
Course Information
Data, homework, lecture notes, etc. for the course will be posted at the Course Home Page.
Textbooks
There are two required texts for this course.
The first required text is Artificial Intelligence: A Modern Approach by Stuart Russell and Peter Norvig, ISBN: 0131038052. If not available in the bookstore, you can get it at amazon.com or Barnes and Noble.
As we will be coding in Lisp, the other text is ANSI Common Lisp by Paul Graham, ISBN: 0133708756. If not available in the bookstore, you can get it at amazon.com or Barnes and Noble. In the past this second text was recommended but not required. That is not the case any more. It is required.
| | | Sec 001 | Sec 002 | Tentative Topic |
| Aug | 27 | 26 | Lisp, Philosophy |
| Sep | 3 | 9 | Simple Systems, Robotics |
| | 10 | 16 | Lisp, Basic Neural Networks |
| | 17 | 23 | Lisp, Backpropagation |
| | 24 | 30 | Lisp, Other Networks |
| Oct | 1 | 7 | Decision Trees |
| | 8 | 15 (not 14!) | Optimization, Evolutionary Computation |
| | 22 | 21 | Midterm |
| | 29 | 28 | Reinforcement Learning |
| Nov | 5 | 4 | History, Uninformed Search |
| | 12 | 11 | Heuristic Search |
| | 19 | 18 | Game Search |
| | 26 | 25 | Knowledge Representation |
| Dec | 3 | 2 | Knowledge Representation |
| | 10 | 16 | Final (not comprehensive) |
Tentative Schedule and Section Migration
The course will be set up so that both sections have identical lectures. This means that you are free to attend one, or the other section, or both, with the constraint that if there are not enough desks, then the desks go to the students signed up for that section. I anticipate that both sections will also have the same TA, as well as the same homework and projects.
However, the sections will have very different midterms and exams, because I cannot give the exams at exactly the same time. This means that you must take the midterm/exam with your section.
Grading Policies
This course will consist of homework and projects, a midterm, and a final exam. The breakdown will be approximately:
| Homework and Projects | 50% |
| Midterm | 25% |
| Final | 25% |
There will be no make-up tests for missed examinations. Late projects will be accepted but at a loss of 20% per day (projects later than 4 days, or beyond the final exam of Section 001 (December 10), will not be accepted). The final will not be comprehensive. The course is graded, more or less, on a curve: this means that in addition to scores for homework, projects, and the midterm, you will also receive a breakdown of the class distribution for that item.