Académique Documents
Professionnel Documents
Culture Documents
03/05/2012
Week 27
Reading List
03/05/2012 3
Introduction
In this lecture, we cover an introduction to KBS. We start with identifying the different types of AI: numerical and symbolic. We look at some search algorithms as simple AI system.
03/05/2012
Topics of Discussion
AI Simple AI systems Developing KBS Some famous KBS
03/05/2012
AI
Artificial Intelligence is the field of computing that attempts at providing computational models of some human activities, which researchers consider intelligent activities, such as learning, acting, decision making, evolving and so on. AI, therefore, relates strongly to fields such as psychology, biology and sociology. In some cases new disciplines emerged such as bio-informatics and cybernetics.
03/05/2012 6
AI
There are two main streams in developing AI systems: quantitive and qualitative approaches. Quantitive approaches sometimes referred to as numerical approaches, because they use quantities in analysing the problems. Neural nets, fuzzy logic, genetic algorithms are all examples of the quantitive approach.
03/05/2012
AI
Qualitative approaches sometimes referred to as symbolic approaches, because they use qualities of the problem to solve the problem. Logic, rules, lists based systems are examples of qualitative AI systems.
03/05/2012
Simple AI systems
The simplest view of AI systems is as a search problem solver. It is almost impossible to develop an expert system without implementing some search technique or another to navigate through the problem domain for the solution. Search techniques provide the base for the inference engine, which is an essential component of any expert system.
03/05/2012 9
Simple AI systems
There are two main types of searches: Conventional searches and heuristic searches. Conventional searches cover the entire domain and eventually find the solution, what is the problem with that? Heuristic searches aim at reducing the domain or covering a selected portion of the problem domain. What is the problem with that?
03/05/2012 10
Simple AI systems
Conventional searches include:
Generate and test. Hill climbing. Best first. Problem reduction. Constraint satisfaction. Means-end analysis.
03/05/2012
11
Developing KBS
(Please refer to the second lecture and lecture notes part 2) Many KBSs are symbolic systems. There are two distinctive parts need to be included in any KBS:
Knowledge representation, which is usually the result of knowledge acquisition Inference Engine, which you would not usually need to develop if you are using an expert system shell such as CLIPS
03/05/2012
12
Developing KBS
In KBS, we also call them exact systems, we do not need to imply certainty factor as we did in FLS. In CLIPS, KBS can be developed as pure rules without the need to define fuzzy sets, i.e. no deftemplate is required.
CLIPS is a productive development and delivery expert system tool which provides a complete environment for the construction of rule and/or object based expert systems., CLIPS was created in 1985 and is now widely used throughout the government, industry, and academia. For further details including its key features, please see http://www.ghg.net/clips/WhatIsCLIPS.html
03/05/2012 13
03/05/2012
14
DENDRAL (1965-83)
DENDRAL (1965-83): The DENDRAL Project was one of the earliest expert systems. DENDRAL began as an effort to explore the mechanization of scientific reasoning and the formalization of scientific knowledge by working within a specific domain of science, organic chemistry. Another concern was to use AI methodology to understand better some fundamental questions in the philosophy of science, including the process by which explanatory hypotheses are discovered or judged adequate. After more than a decade of collaboration among chemists, geneticists, and computer scientists, DENDRAL had become not only a successful demonstration of the power of rule-based expert systems but also a significant tool for molecular structure analysis, in use in both academic and industrial research labs. Using a plangenerate-test search paradigm and data from mass spectrometry and other sources, DENDRAL proposes plausible candidate structures for new or unknown chemical compounds. Its performance rivals that of human experts for certain classes of organic compounds and has resulted in a number of papers that were published in the chemical literature. Although no longer a topic of academic research, the most recent version of the interactive structure generator, GENOA, has been licensed by Stanford University for commercial use. (taken from http://smi-web.stanford.edu/projects/history.html)
15
03/05/2012
MYCIN (1972-80)
MYCIN is an interactive program that diagnoses certain infectious diseases, prescribes antimicrobial therapy, and can explain its reasoning in detail. In a controlled test, its performance equalled that of specialists. In addition, the MYCIN program incorporated several important AI developments. MYCIN extended the notion that the knowledge base should be separate from the inference engine, and its rule-based inference engine was built on a backwardchaining, or goal-directed, control strategy. Since it was designed as a consultant for physicians, MYCIN was given the ability to explain both its line of reasoning and its knowledge. Because of the rapid pace of developments in medicine, the knowledge base was designed for easy augmentation. And because medical diagnosis often involves a degree of uncertainty, MYCIN's rules incorporated certainty factors to indicate the importance (i.e., likelihood and risk) of a conclusion. Although MYCIN was never used routinely by physicians, it has substantially influenced other AI research. At the HPP, MYCIN led to work in TEIRESIAS, EMYCIN, PUFF, CENTAUR, VM, GUIDON, and SACON, all described below, and to ONCOCIN and ROGET. The book Rule-Based Expert Sytem: The MYCIN Experiment at the Stanford Heuristic Programming Project describes the decade of research on MYCIN and its descendants. (taken from http://smiweb.stanford.edu/projects/history.html)
03/05/2012 16
R1/XCON (1980s)
One of the first commercially successful expert systems
Application domain: configuration of minicomputer systems selection of components arrangement of components into modules and cases Approach data-driven, forward chaining consists of about 10,000 rules written in OPS5 Results quality of solutions similar to or better than human experts roughly ten times faster (2 vs. 25 minutes) estimated savings $25 million/year
03/05/2012 17
Conclusion
AI systems and search algorithms. Developing KBS.
03/05/2012
18
Next Steps
Next
Knowledge acquisition.
03/05/2012
19