Expert Systems: Human Issues (Artificial Intelligence) - Tapa dura

Reseña del editor

If a computer is providing "advice" - perhaps relating to questions that directly effect people's health or well-being - how far can that advice be trusted, and who is held responsible if it turns out to be wrong? Do the rifles that human experts believe they apply encapsulate all their knowledge, or do their decisions depend upon other, implicit, knowledge that cannot be captured in a computer program? This book brings together firsthand reports from people who have been confronting such important human issues in a number of working or prototype systems in domains that range from child abuse cases and social security claims to an industrial system used in a bonding shop.

The lessons to be drawn are clear: human factors must be fully considered right from the start. They can be decisive in determining not only whether the completed system fulfills the role it was designed for but also whether an expert system is likely to be a practicable proposition in the a particular context.

In four parts, the book describes case studies (DISPLAN - designing a usable medical expert system, user acceptance of a knowledge-based system for the management of child abuse cases, and examples of knowledge engineering in a difficult domain), methods for system development (the application of Kads methodology in the ICL Kidsgrove Bonding Shop, issues in the design of expert systems for business, and general methods for the development and application of knowledge based systems), effective interfaces (connection diagrams, multimodal communication graph implementation), and consequences for individuals and organizations.

Dianne Berry is a Research Fellow in the Department of Experimental Psychology at Oxford University. Anna Hart is a Lecturer in Computer Studies at Lancashire Polytechnic. Expert Systems: Human Issues is included in the Artificial Intelligence Series, edited by Michael Brady, Daniel Bobrow, and Randall Davis.