While we have no evidence, we suspect that most professors write textbooks because they are going to teach a course for which none of the existing texts are suitable. Whether or not this is typically the case, it was certainly the reason for the book you are now holding. There were three things we wanted from a text on Artificial Intelligence (from now on "Al").

It had to include those aspects of the field that we felt would prove to be enduring (for example, low-level vision research, which is given short shrift in most Al textbooks).

It should present Al as a coherent body of ideas and methods, drawing on the work of the last twenty years, but translating it into a more uniform set of concepts, favoring clarity over journalistic accuracy.

It should acquaint the student not only with the classic programs in the field, but the underlying theory as well: mathematical logic for knowledge representation and inference, linguistics for language comprehension, etc.

In brief, our belief was that Al is distinguished by its subject matter, not its his- tory or methods. Its subject matter is the mind, considered as an information processing system.

When each of us became aware that the other was thinking along these lines, we decided to pool our efforts and write a textbook to reflect our view of the field. Four years later, here is the result.

Writing a textbook turned out to be a good way to develop humility about the shortcomings of other textbooks. For one thing, a general-purpose textbook must cover the significant results in all aspects of the field, including some its authors know little about. So we had to do some reading and writing we hadn't anticipated.

No other version available

No Data