The course will be organized around the enterprise of building an artificial rational agent. We will begin by building a very simple agent, and then we will make it progressively more sophisticated over the course of the semester. We will focus primarily on automated reasoning and automated planning. The objective is to construct an agent that can acquire information through (simulated) perception, reason about how to achieve its goals, and act on the basis of its deliberations.
A knowledge of formal logic equivalent to that gained by taking an introduction to symbolic logic (Phil/Math 202) will be assumed. Each year some students find themselves in difficulty towards the end of the semester because they have not taken this requirement seriously. To see whether you know the requisite material, or to learn more about symbolic logic, go the the website for Phil 202. You can download an ftp version of the text. You will be assumed to know at least the material in the first four chapters.
Some familiarity with computer programming will also be assumed. One of the objectives of the course will be to teach the students the rudiments of AI programming in the LISP programming language. I assume that most of the students have no previous familiarity with LISP.
The main text is Artificial Intelligence--A Modern Approach, by Stuart Russell and Peter Norvig. This is generally regarded as the best textbook in AI currently available. However, it is expensive, and we will only have time to talk about the material in a small part of the book. You may want to adjust your book purchasing strategies with this in mind,
The LISP text is Common LISPcraft, by Wilensky.
There will be a number of exercises that will require some programming in LISP. On most Fridays, the course will meet in the Instructional Computing Lab in Soc. Sci. 226, and that time will be devoted to learning programming and going over the exercises. The lab meetings will be run by Josh ..., who is the TA for the course. The university has acquired a site license for Allegro Common LISP, which runs in Windows. This has been installed on the computers in the Instructional Computing Lab. It will also be made available to you for use on your own machine. Some provision will also be made for Mac users.
The course will be graded entirely on the basis of the exercises.
Teaching assistant:Office hours:
Pollock: Monday 1:00 to 2:00 and Friday 1:30 - 2:00 or by appointment.
Josh: Monday and Wednesday, 3:00 - 4:00, or by appointment.
These are to be turned in by email to Josh at one hour before class on the due date. No late homework will be accepted under any circumstances.
The assignments are listed on a separate website maintained by Josh.