Artificial Intelligence

Lecturer: Dr W.F. Clocksin (wfc@cl.cam.ac.uk)

No. of lectures: 12

Prerequisite course: Prolog for Artificial Intelligence

Aims

The aim of this course is to introduce the principles and applications of those aspects of artificial intelligence that are not covered by other courses. Other courses relevant to AI are: Prolog (from Part IB), Computer Vision, Neural Computing, Information Retrieval, and Natural Language Processing. This course is more about issues and ideas than about problems and their solutions. No advanced mathematics is required, only basic programming experience and a working knowledge of Prolog. For these reasons, there will be four unassessed Examples Classes instead of supervisions. The Examples Classes will work through typical examination questions and discuss issues of interest to the participants.

Topics

• Problems and challenges. Why is something that looks so easy to do actually so difficult to compute?

• Theories and methods. What approaches have been tried?

• AI programming. What is LISP for, and how does it relate to Prolog?

• Constraints. How can one exploit the natural structure of a problem in order to solve it?

• Knowledge representation. What are the problems with representing commonsense knowledge?

• Reasoning. What formal methods are there for dealing with ill-posed problems?

• Inference. What follows from what? How do you know it does?

• Fuzzy logic. How can one compute and control using approximations in the absence of a model of the process?

• Qualitative control. How little precision can you get away with?

• Search. What are some ways of looking for something?

• Planning. How can you figure out what to do before you do it?

• Decision making. What is a basis for making a decision, and why does it matter?

Objectives

At the end of the course students should

• appreciate the distinction between the popular view of the field and the actual research results

• appreciate different perspectives on what the problems of artificial intelligence are and how different approaches are justified

• be able to design problem-solving methods based on search and reasoning techniques

• know how to model situations and actions using a variety of knowledge representation techniques

Recommended books

Luger, G.F. & Stubblefield, W.A. (1998). Artificial Intelligence: Structures and Strategies for Complex Problem Solving. Addison-Wesley.
Russell, S. & Norvig, P. (1995). Artificial Intelligence: A Modern Approach. Prentice-Hall.
Shoham, Y. (1994). Artificial Intelligence Techniques in Prolog. Morgan Kaufmann.