Principles of Expert Systems
Peter Lucas, Linda van der Gaag · 1991
(copyright returned to the authors) iiHistory of the Book This book was written from 1989 to 1990 and it took a year before it was ready for publication by Addison-Wesley. From a modern AI point of view, it gives a detailed account of methods and techniques, often including details of their implementation, of the systems that caused a hype during the 1980s. At the time, there were books that gave some insight into methods, but none of the books gave actual details on algorithms and implementation. Thus the book was written to close the gap between theory (and there is plenty of theory in this book) and implementation, making theoretical ideas concrete. Unfortunately, when the book was published, the AI winter had started and it never became a success. From a modern point of view, this book is now one of few books that give details about methods and systems that once were in the centre of interest. It establishes relationships between logic and rule-based methods which you will find in no other books. The chapter on reasoning with uncertainty is also still of interest. It also gives details about many of the early AI methods, languages and systems. Although the title of the book mentions the phrase ‘Expert Systems’, the book is in reality about knowledge (based) systems. Nevertheless, it is assumed that such systems are built using human knowledge only. Since the beginning of the 1990s there has been considerable emphasis on using machine learning methods to build such systems automatically. However, now we know that machine learning also has its limitations and that you cannot build such systems without a considerable amount of human knowledge.