The representation of uncertainty is a central issue in Artificial Intelligence (AI) and is being addressed in many different ways. Each approach has its proponents, and each has had its detractors. However, there is now an in creasing move towards the belief that an eclectic approach is required to represent and reason under the many facets of uncertainty. We believe that the time is ripe for a wide ranging, yet accessible, survey of the main for malisms. In this book, we offer a broad perspective on uncertainty and approach es to managing uncertainty. Rather than provide a daunting mass of techni cal detail, we have focused on the foundations and intuitions behind the various schools. The aim has been to present in one volume an overview of the major issues and decisions to be made in representing uncertain knowl edge. We identify the central role of managing uncertainty to AI and Expert Systems, and provide a comprehensive introduction to the different aspects of uncertainty. We then describe the rationales, advantages and limitations of the major approaches that have been taken, using illustrative examples. The book ends with a review of the lessons learned and current research di rections in the field. The intended readership will include researchers and practitioners in volved in the design and implementation of Decision Support Systems, Ex pert Systems, other Knowledge-Based Systems and in Cognitive Science.
This text identifies the central role of managing uncertainty in AI and expert systems and provides a comprehensive introduction to different aspects of uncertainty and the rationales, descriptions (through worked examples), advantages and limitations of the major approaches that have been taken. The book introduces and describes the main ways in which uncertainty can occur and the importance of managing uncertainty for the production of intelligent behaviour in AI and its associated technologies of knowledge-based systems. It also describes the rationale, advantages and limitations of the major representational approaches (both quantitative and symbolic) that have been employed in AI systems and provides a worked illustration of each method. Finally, the book summarizes the significant themes that have emerged from applications and the research literature and identifies current and future directions. Concentrating wholly on this specific area of Artificial Intelligence, the work is aimed primarily at researchers and practitioners involved in the design and implementation of expert systems, other knowledge-based systems and cognitive science.
It should also be of value to students of computer science, cognitive science, psychology and engineering with an interest in AI or decision support systems. While a technical book, technical details are presented in appendices, allowing the text to be read continuously by nontechnical readers.
