Browse Books

Reviewer: Hsun Chang

This well-written book is a wonderful text for researchers working on theoretical artificial intelligence (AI). Fundamentally, AI represents knowledge with mathematical objects and then designs computational rules to manipulate these objects. Among many, the graph-based representation scheme is popular, because of its effectiveness in capturing knowledge and its computability for object manipulation. Chapter 1 opens the book with an introduction. Although this book is mostly theory oriented, this chapter gives the big picture of knowledge representation and describes the uniqueness of the graph-based approach. Furthermore, an example is used to illustrate how graphs can be useful in an AI system. Despite being succinct, this chapter is a must for the reader unfamiliar with knowledge representation. Part 1 consists of chapters 2 to 5. This part mathematically defines objects in the graph-based framework. Conceptual graphs, which are the core of this framework, can describe entities and their relations. For example, "Mary is a girl" is described by a graph where a node denotes "girl," another node represents "Mary," and the link between the two nodes specifies that "Mary is a girl." Obviously, when a piece of knowledge is complicated, the corresponding graph representation also becomes complex. Through a systematic way of defining knowledge representation, we can decompose the knowledge into pieces of graphs with distinct properties. The second part is composed of chapters 6 to 8. After learning the formal definitions of graphical representations introduced in the preceding part, we can utilize computational methods to explore the properties of the graphs. For example, we may want to compare a given graph with another in our knowledge database, or we may reason and infer knowledge beyond the information that has been captured by graphs. This part details graph operation algorithms, each of which is equipped with pseudocode, so the reader can readily apply the techniques in practice. The remaining chapters constitute Part 3, which extends the primaries to more complicated cases. For example, a graph that shares the same structure with its subgraph-which is imagined as a node in the graph-is a nested graph, commonly occurring in knowledge representation. Although the methods presented in this part are better for describing knowledge in the real world, they require more sophisticated mathematics, so these chapters involve more theorems and proofs. This book assumes that the reader is familiar with fundamental graph theory. Although the appendix briefly presents the bare necessities of graph theory required to understand this book, it is better to learn graph theory before reading the book. Otherwise, you will be frustrated in the middle, particularly if you are an entry-level student. In summary, this is a theoretical book for a graph-based approach to knowledge representation. It presents it rigorously, from bottom to top. A number of detailed algorithms presented in the book may serve as good references for designing a variety of AI systems, such as database mining and logic reasoning. Online Computing Reviews Service