|
In December 1998, Addison Wesley
Longman brought out the 3rd Edition of my expert systems text. This book
has been rather a long time in the making, thanks to various relocations,
career changes, health problems, and a great deal of goofing off on my
part.
After writing the 2nd Edition, I quit defense computing in
1992 and returned to academia to teach Expert Systems for the first time
since leaving Edinburgh, as well as Artificial Intelligence, Numeric
Computation, Symbolic Computation and Parallel Systems. I also taught
classes on AI, Expert Systems, CLIPS and Scheme Programming at Singapore
Polytechnic in the summer of 1994. These experiences allowed me to
troubleshoot the 2nd Edition to some extent, and gave me ideas for
improving the pedagogical aspects of the book next time
around.
 The 1st Edition, published in 1986, was arguably the
first text written primarily for students, as opposed to researchers
or the business world. |
 The 2nd Edition added chapters on more advanced
topics, such as blackboard systems,
machine learning, and truth maintenance. |
Features of the New
Edition
The 3rd Edition consolidates the treatment of the previous
edition in a number of ways.
It provides a gentler introduction to AI and expert system
concepts with more informal examples, many of them contained in sidebars
or boxes beside the main text.
It features chapters that introduce new concepts, such as case
based reasoning, connectionism, explanation-based generalization,
apprenticeship learning, and hybrid systems.
It features new systems and frameworks, such as COMPASS, CATO,
CHEF, KADS, ODYSSEUS, FRANK, C4.5, SCALIR, and GBB.
It provides a substantial Appendix on CLIPS
programming.
CLIPS is used as the main programming language throughout the
text, including examples that feature the CLIPS Object-Oriented Language,
COOL. |
|
Chapter
Outline of the New Edition
Here is the chapter outline of the 3rd Edition. Many
pre-existing chapters have been substantially rewritten, while some of the
later chapters contain all new material. The text is much more nonlinear
than the 2nd Edition, with an emphasis on ideas rather than systems.
Sidebars introduce fresh material which ties expert system concepts more
closely to fundamental topics in artificial intelligence and computer
science.
1. What Are Expert Systems?
2.
An Overview of Artificial Intelligence
3. Knowledge
Representation
4. Symbolic Computation
5. Rule-Based
Systems
6. Structured Objects
7. Object-Oriented
Programming
8. Logic Programming
9. Representing
Uncertainty
10. Knowledge Acquisition
11. Heuristic
Classification I
12. Heuristic Classification
II |
13. Hierarchical Hypothesize &
Test
14. Constructive Problem Solving I
15. Constructive
Problem Solving II
16. Designing for Explanation
17. Tools
for Building Expert Systems
18. Blackboard Systems
19. Truth
Maintenance Systems
20. Explorations in Machine Learning
21.
Belief Networks
22. Case Based Reasoning
23. Hybrid
Systems
24. Summary and
Conclusion |
The chapter
on hybrid systems introduces some '90s expert systems that have a learning
component, e.g., combining case based reasoning, explanation based
generalization, or connectionism with more traditional rule based
approaches.
The 3rd Edition has more programming exercises than the previous editions. The CLIPS language is used throughout as the main study
vehicle to provide continuity, but there are also some examples in LISP,
PROLOG and C++. The book ends with a 50-page Appendix on CLIPS which
contains some nontrivial programs.
|
