Version of 7 January 2006
This page gives you an overview of the courses - how they will be organized, what my responsibilities are, and what your responsibilities are. More importantly it gives you a quick overview of the importance, pervasiveness, and excitement of the field of Artificial Intelligence.
The course will begin with overview lectures and readings covering every chapter in our textbook. The readings for this will cover only a small portion of each of the 27 chapters, typically the first one or two sections of each chapter as well as the chapter summary (usually one page). There are a number of reasons for starting the course out this way, including: Giving you a "birds eye view of the field"; Allowing you to think about the topic for your course project without waiting for many weeks for a potentially interesting topic to be discussed; Exposing you to the material you'll later be responsible for so you can see how well you are or are not prepared for it; Discovering topics that you can begin reading about in more detail, both in the textbook and in the many books and articles that will be on Reserve.
The other important component of the course, besides the usual lectures and exams, will be your semester project. Your project material will be handed in in three stages, plus a penultimate handin that will allow me to give you my final suggestions before the completed project is handed in. The great majority of the projects can and should require you to do some programming. Further details are on the Projects page.
My responsibilities include giving you all manner of course information in a timely manner, answering your questions in and out of class, telling you what is expected on exams and your projects, and of course lecturing. I'm always happy to respond to questions before, during, and after class. If the dialogue becomes too complex or drawn out, we can take it offline.
Read this brief note on your responsibilities first. Your primary responsibilities are to yourself. You owe it to yourself to learn as much as you can about this fascinating and challenging material. Part of your responsibilities include coming to class. Much of the material needs explaining beyond the textbook, which can move quickly and in non-obvious ways. So keep up with the course. This is not a course in which you want to get behind, especially during the full course Overview at the beginning, which you will be tested on. You are responsible for reading your mail and for keeping in touch with questions and explanations of what you're doing, especially if there are complications. I may send mail at any time of the night or day, so it's a good idea to check your mail a couple of times a day. You are responsible for reading the assigned sections of your textbook. Don't read passively, especially when studying for an exam. Taking an exam consists of reading the questions carefully and then writing, so always write while studying. (Highliting is a weak substitute for writing.) Your writing/notes for this course should contain a mix of prose, diagrams, and math/code.
AI has matured to the point that it has become embedded in many important systems. The speech recognition system that you encounter when phoning businesses were originally difficult AI problems. But they were solved quite well and now are commercial products. The language translation systems on sites such as Google are far beyond word-for-word translations and owe their success to years of AI-based R&D. (The Director of Research at Google is Peter Norvig, one of the authors of your textbook.) Huge data mining systems such as the ones used by retailers such as WalMart and Amazon all grew out of AI research on machine learning. The route finding systems behind MapQuest and Google maps, and the route planning behind UPS and FedEx, are all outgrowths of AI work on these problems. The challenges that AI faces today in building larger and smarter systems in new areas are huge. AI will be with us forever - the problem of replicating human intelligence and building systems that go beyond human intelligence in certain ways will challenge us for the indefinite future.
There are "standard" approaches to AI courses. They reflect the contents of AI textbooks, such as the excellent text used in this course or the one by Nilsson that is on reserve for the course. But I will approach AI from a somewhat different perspective. I will supplement the textbook topics with excursions into Cognitive Science and Game AI. Cognitive Science focuses on human cognition, the very thing that AI is attempting to emulate. So it is important to understand something about human intelligence. This may sound like a trivial problem - we're all human so we must know how we perceive, think, and solve problems.
But much of what we would like to believe about ourselves has turned out, on the basis of many careful experiments, to be far different than we would imagine. People are systematically illogical and systematically unable to estimate the quantitative aspects of things. Perception is similar. Simply printing the word "blue" in red reveals interactions between the visual and language centers. When reading, the characters only six or eight characters to the left or right of the ones you are focused on can be changed arbitrarily without you noticing it. During a blink, major items in a scene can be altered without you being aware of it when your eyes open again. These types of studies, and much more, are discussed at length in the five books on Cognitive Science on reserve for the course.
There will also be some discussions of animal intelligence, a phrase that was an oxymoron only a few decades ago. Birds can count up to six or eight, and memorize the locations of thousands of food items stored in the ground for later use. The list of strange and wonderful abilities and incapabilities of humans and other animals goes on and on. There are two books on reserve for the course that are specifically focused on animal intelligence. They make fascinating reasoning.
There are strong requirements in computer and console games for the use of Artifical Intelligence techniques. This revolves around the design of non-player characters (NPCs). If the characters in a game that you are collaborating with or trying to outwit or defeat are "stupid", then the game quickly becomes boring. I recently purchased an XBOX 360 and some games. I will do some game play in class to demonstrate how AI is integrated into modern games. Initially I got Gears of War and Oblivion. This is new territory for me - we'll see how it goes. There are four game AI books on reserve for this course, each with their own emphasis. Enjoy!
Cognitive Science, animal intelligence, and games offer useful insights into the problems of building artificially intelligent systems.
Some material that complements the notes above can be found in the introduction to my Spring 2006 AI courses.
Go to CSU520 home page.,
to CSG120 home page.,
RPF's Teaching Gateway, or
homepage