1

2 Inexact Reasoning 2 Session 10Course : T0273 – EXPERT SYSTEMS Year : 2014 Inexact Reasoning 2 Session 10

3 Learning Outcomes LO 3 : Solve problems by using Expert SystemsAfter taking this course, students should be expected to understand and explain about inexact reasoning. T Expert Systems

4 Lecture Outline Approximate Reasoning The State of UncertaintySome Commercial Applications of Fuzzy Logic Summary Exercise T Expert Systems

5 Approximate ReasoningA theory of uncertainty based on fuzzy logic is primarily concerned with quantifying and reasoning using natural language in which many words have ambiguous meanings. As we all know, sleeping a little longer is a very inexact term that varies from person to person. The more complex something is, the more inexact or “fuzzier” it will be. Fuzzy logic provides a precise approach for dealing with uncertainty which grows out of the complexity of human behavior. T Expert Systems

6 Approximate ReasoningThe term soft computing has come to mean computing that is not based on the classical two-valued logics. Soft computing includes fuzzy logic, neural networks, and probabilistic reasoning. The theory has been extended and applied to many fields for a long time, such as automatic camera tracking of an object in space. Fuzzy logic has also been combined with neural networks in many applications. T Expert Systems

7 Approximate ReasoningSome applications of fuzzy theory: Control Algorithms Medical Diagnosis Decision Making Economics Engineering Environmental Literature Operations Research Pattern Recognition Psychology Reliability Security Science T Expert Systems

8 Approximate ReasoningTranslation rules Fuzzy probability is incorporated into the fuzzy logic called FL. One main component of FL is a group of translation rules. Type I: modification rules Type II: composition rules conditional composition conjunctive composition disjunctive composition conditional and conjunctive composition Type III: quantification rules Type IV: qualification rules truth qualification probability qualification possibility qualification T Expert Systems

9 The State of UncertaintyThere are two mountains that stand out from all the trees and forests: Mountain of Logic Mountain of Uncertainty An expert system must come up with valid conclusions given that: The rules were written correctly. The facts on which the inference engine generates valid conclusions are true facts. T Expert Systems

10 The State of UncertaintyMountain of Uncertainty The best we can do on this mountain is model it on the expertise of our expert, or try to include more than one approach to uncertainty and let the different techniques fight it out. Today fuzzy logic and Bayesian theory are most often used for uncertainty. T Expert Systems

11 Some Commercial Applications of Fuzzy LogicMany commercial applications of fuzzy logic are in everything from cameras to washing machines: Hydroelectric-powerplants Robots Camera aiming Assessment of stock exchange activities Air-conditioning systems Car-engines Automobiles Industrial control applications Production of semiconductors Bus time-tables etc. T Expert Systems

12 Summary Fuzzy logic provides a precise approach for dealing with uncertainty which grows out of the complexity of human behavior. Soft computing includes fuzzy logic, neural networks, and probabilistic reasoning. Fuzzy probability is incorporated into the fuzzy logic called FL. There are two mountains that stand out from all the trees and forests: Mountain of Logic and Mountain of Uncertainty T Expert Systems

13 Exercise Answer the questions below:Define at least six values for the linguistic variable Water Temperature. Draw the appropriate functions for the fuzzy set values on one graph. Given numeric truth values, x(A) = .2/ /.5 + 1/.9 x(B) = .1/ /.5 + 1/.9 calculate the fuzzy logic truth of the following: a) NOT A d) A  B b) A AND B e) B  A c) A OR B T Expert Systems

14 References Joseph Giarratano, Gary Riley Expert Systems: Principles and Programming Chapter 5. Thomson Course Technology. Australia. ISBN: Peter Jackson Introduction to Expert Systems. Addison-Wesley. Harlow, England. ISBN: T Expert Systems