One of the great challenges facing educators today is the increasing dependence of different disciplines on one another and the difficulty of equipping students in one discipline with the tools in other areas that they will need to function effectively. This book attempts to bridge the domains of library science and artificial intelligence (AI) by offering the undergraduate student of library science an introduction to AI. The first chapter sets the context by examining the concept of knowledge and distinguishing factual knowledge (the traditional domain of library science) from inference knowledge (the concern of AI). The second chapter then reviews factual knowledge and how it can be organized and classified. The following five chapters introduce state-space search, extensions of search using heuristic evaluation functions, symbolic inference, the use of semantic networks and scripts in supporting dialogue, and finally an overview of expert systems. Thus the book is designed to enable a librarian to engage expert systems and their underlying mechanisms with some degree of fluency. The preface promises a second volume to cover other topics in AI that this book omits, including qualitative reasoning, knowledge acquisition, and analogical reasoning. Though the book is a translation from a Japanese original, the English is fluent and simple. Succinct summaries, keyword lists, and exercises for each chapter make the book appropriate either for self-study or for classroom use. The book is short enough to cover in a single semester.

When introducing a technical specialty such as AI to those who never intend to specialize in it, one risks either losing the students in technical detail or oversimplifying the subject matter. This volume is more in jeopardy of losing the student, largely because of its selection of examples. After the first two chapters, these have little or nothing to do with library science, but are drawn from domains--such as calculus and formal linguistics--that are likely to be much more familiar to budding AI scientists than to librarians in training. The book would serve its intended audience much better if it were reworked to illustrate the otherwise clear exposition with examples from a plausible application of AI technology to library science. In its present form, one questions whether it offers a significant advantage over a standard introduction to AI such as Winston [1].

The book concentrates on elementary issues, but contains brief introductions to a wide range of related topics. For example, the chapter on logical inference concentrates on straightforward first-order predicate calculus, but includes brief sections on such topics as the closed-world assumption, default-inference, non-monotonic logics, commonsense reasoning, and circumscription. A careful reader will come away with a solid backbone of material and some useful hooks for future extension.

The book is attractively prepared, with copious graphics, appropriate use of various type fonts, and a reasonable index. A few unfortunate lapses occur in the proofreading, such as the spelling of “knowledge” in the title of chapter 6 in the table of contents, the variation between “inference knowledge” in chapter 1 and “inferential knowledge” in later chapters, and the failure to set off the heading to section 5.7.5. The translation is in general smooth and idiomatic, but could be improved if Japanese personal names (such as “Taro,” “Hanako,” and “Jiro”) in examples were replaced with Western personal names.