Expert Systems

Department of Information
Technology
Ambo University
 What is an expert system?

“An expert system is a computer system that

emulates, or acts in all respects, with the
decision-making capabilities of a human expert.”

Professor Edward Feigenbaum

Stanford University

2
Areas of Artificial Intelligence

3
 Expert system technology

 Special expert system languages – CLIPS

 Programs

 Hardware designed to facilitate the

implementation of those systems

4
 Expert System Main Components

 Knowledge base – obtainable from books,

magazines, knowledgeable persons, etc.

 Inference engine – draws conclusions from the

knowledge base

5
Basic Functions of Expert Systems

6
Problem Domain vs. Knowledge Domain

 An expert’s knowledge is specific to one problem

domain –
 medicine, finance, science, engineering, etc.
 The expert’s knowledge about solving specific
problems is called the knowledge domain.
 The problem domain is always a superset of the
knowledge domain.

7
Problem and Knowledge Domain Relationship

8
 Advantages of Expert Systems

 Increased availability
 Reduced cost
 Reduced danger
 Performance
 Multiple expertise
 Increased reliability

9
 Advantages Continued

 Explanation

 Fast response

 Steady, unemotional, and complete responses at

all times

 Intelligent tutor

 Intelligent database

10
 Representing the Knowledge

The knowledge of an expert system can be

represented in a number of ways, including IF-
THEN rules:

IF you are hungry THEN eat

11
 Knowledge Engineering

The process of building an expert system:

1. The knowledge engineer establishes a dialog

with the human expert to elicit knowledge.
2. The knowledge engineer codes the knowledge
explicitly in the knowledge base.
3. The expert evaluates the expert system and
gives a critique to the knowledge engineer.

12
Development of an Expert System

13
 The Role of AI

 An algorithm is an ideal solution guaranteed to

yield a solution in a finite amount of time.
 When an algorithm is not available or is
insufficient, we rely on artificial intelligence
(AI).
 Expert system relies on inference – we accept a
“reasonable solution.”

14
 Uncertainty

 Both human experts and expert systems must be

able to deal with uncertainty.
 It is easier to program expert systems with
shallow knowledge than with deep knowledge.
 Shallow knowledge – based on empirical and
heuristic knowledge.
 Deep knowledge – based on basic structure,
function, and behavior of objects.

15
 Limitations of Expert Systems

 Typical expert systems cannot generalize through

analogy to reason about new situations in the way
people can.

 A knowledge acquisition bottleneck results from

the time-consuming and labor intensive task of
building an expert system.

16
 Early Expert Systems

 DENDRAL – used in chemical mass

spectroscopy to identify chemical constituents
 MYCIN – medical diagnosis of illness
 DIPMETER – geological data analysis for oil
 PROSPECTOR – geological data analysis for
minerals
 XCON/R1 – configuring computer systems

17
Broad Classes of Expert Systems

18
 Problems with Algorithmic Solutions

 Conventional computer programs generally solve

problems having algorithmic solutions.

 Algorithmic languages include C, Java, and C#.

 Classic AI languages include LISP and

PROLOG.

19
Considerations for Building Expert Systems

 Can the problem be solved effectively by

conventional programming?
 Is there a need and a desire for an expert system?
 Is there at least one human expert who is willing
to cooperate?
 Can the expert explain the knowledge to the
knowledge engineer can understand it.
 Is the problem-solving knowledge mainly
heuristic and uncertain?

20
 Languages, Shells, and Tools

 Expert system languages are post-third

generation.
 Procedural languages (e.g., C) focus on
techniques to represent data.
 More modern languages (e.g., Java) focus on data
abstraction.
 Expert system languages (e.g. CLIPS) focus on
ways to represent knowledge.

21
 Elements of an Expert System

 User interface – mechanism by which user and

system communicate.
 Exploration facility – explains reasoning of
expert system to user.
 Working memory – global database of facts
used by rules.
 Inference engine – makes inferences deciding
which rules are satisfied and prioritizing.

22
 Elements Continued

 Agenda – a prioritized list of rules created by the

inference engine, whose patterns are satisfied by
facts or objects in working memory.
 Knowledge acquisition facility – automatic way
for the user to enter knowledge in the system
bypassing the explicit coding by knowledge
engineer.

23
 Production Rules

 Knowledge base is also called production

memory.

 Production rules can be expressed in IF-THEN

pseudocode format.

 In rule-based systems, the inference engine

determines which rule antecedents are satisfied
by the facts.

24
 Structure of a
Rule-Based Expert System

25
 General Methods of Inferencing

 Forward chaining – reasoning from facts to the

conclusions resulting from those facts – best for
prognosis, monitoring, and control.

 Backward chaining – reasoning in reverse from a

hypothesis, a potential conclusion to be proved to
the facts that support the hypothesis – best for
diagnosis problems.

26
 Production Systems

 Rule-based expert systems – most popular type

today.
 Knowledge is represented as multiple rules that
specify what should/not be concluded from
different situations.
 Forward chaining – start w/facts and use rules do
draw conclusions/take actions.
 Backward chaining – start w/hypothesis and look
for rules that allow hypothesis to be proven true.

27
 Forward/Backward Chaining

 Forward chaining – primarily data-driven.

 Backward chaining – primarily goal driven.

28
 Post Production System

 Basic idea – any mathematical / logical system is

simply a set of rules specifying how to change
one string of symbols into another string of
symbols.

 Basic limitation – lack of control mechanism to

guide the application of the rules.

29
 Markov Algorithm

 An ordered group of productions applied in order

or priority to an input string.

 If the highest priority rule is not applicable, we

apply the next, and so on.

 An efficient algorithm for systems with many

rules.

30
 Rete Algorithm

 Functions like a net – holding a lot of

information.
 Much faster response times and rule firings can
occur compared to a large group of IF-THEN
rules which would have to be checked one-by-
one in conventional program.
 Takes advantage of temporal redundancy and
structural similarity.
 Drawback is high memory space requirments.

31
 Procedural Paradigms

 Algorithm – method of solving a problem in a

finite number of steps.

 Procedural programs are also called sequential

programs.

 The programmer specifies exactly how a problem

solution must be coded.

32
Procedural Languages

33
 Imperative Programming

 Focuses on the concept of modifiable store –

variables and assignments.
 During execution, program makes transition from
the initial state to the final state by passing
through series of intermediate states.
 Provide for top-down-design.
 Not efficient for directly implementing expert
systems.

34
 Nonprocedural Paradigms

 Do not depend on the programmer giving exact

details how the program is to be solved.
 Declarative programming – goal is separated
from the method to achieve it.
 Object-oriented programming – partly imperative
and partly declarative – uses objects and methods
that act on those objects.
 Inheritance – (OOP) subclasses derived from
parent classes.

35
Nonprocedural Languages

36
The Representation of Knowledge
 What is the study of logic?

 Logic is the study of making inferences – given a

set of facts, we attempt to reach a true
conclusion.
 An example of informal logic is a courtroom
setting where lawyers make a series of inferences
hoping to convince a jury / judge .
 Formal logic (symbolic logic) is a more rigorous
approach to proving a conclusion to be true /
false.

38
 Why is Logic Important

 We use logic in our everyday lives – “should I

buy this car”, “should I seek medical attention”.
 People are not very good at reasoning because
they often fail to separate word meanings with
the reasoning process itself.
 Semantics refers to the meanings we give to
symbols.

39
 The Goal of Expert Systems

 We need to be able to separate the actual

meanings of words with the reasoning process
itself.
 We need to make inferences w/o relying on
semantics.
 We need to reach valid conclusions based on
facts only.

40
 Knowledge vs. Expert Systems

 Knowledge representation is key to the success

of expert systems.
 Expert systems are designed for knowledge
representation based on rules of logic called
inferences.
 Knowledge affects the development, efficiency,
speed, and maintenance of the system.

41
 Arguments in Logic

 An argument refers to the formal way facts and

rules of inferences are used to reach valid
conclusions.

 The process of reaching valid conclusions is

referred to as logical reasoning.

42
 How is Knowledge Used?

 Knowledge has many meanings – data, facts,

information.
 How do we use knowledge to reach conclusions
or solve problems?
 Heuristics refers to using experience to solve
problems – using precedents.
 Expert systems may have hundreds / thousands
of micro-precedents to refer to.

43
 Epistemology

 Epistemology is the formal study of knowledge .

 Concerned with nature, structure, and origins of

knowledge.

44
Categories of Epistemology

•Philosophy •A priori

•A posteriori •Procedural

•Declarative •Tacit

45
 A Priori Knowledge

 “That which precedes”

 Independent of the senses

 Universally true

 Cannot be denied without contradiction

46
 A Posteriori Knowledge

 “That which follows”

 Derived from the senses

 Now always reliable

 Deniable on the basis of new knowledge w/o

the necessity of contradiction

47
 Procedural Knowledge

Knowing how to do something:

• Fix a watch
• Install a window
• Brush your teeth
• Ride a bicycle

48
 Declarative Knowledge

 Knowledge that something is true or false

 Usually associated with declarative statements

 E.g., “Don’t touch that hot wire.”

49
 Tacit Knowledge

 Unconscious knowledge

 Cannot be expressed by language

 E.g., knowing how to walk, breath, etc.

50
 Knowledge in Rule-Based
Systems
 Knowledge is part of a hierarchy.

 Knowledge refers to rules that are activated by

facts or other rules.
 Activated rules produce new facts or conclusions.

 Conclusions are the end-product of inferences

when done according to formal rules.

51
 Expert Systems vs. Humans

 Expert systems infer – reaching conclusions

as the end product of a chain of steps called
inferencing when done according to formal
rules.

 Humans reason

52
 Expert Systems vs. ANS

 ANS does not make inferences but searches for

underlying patterns.

 Expert systems
o Draw inferences using facts
o Separate data from noise
o Transform data into information
o Transform information into knowledge

53
 Metaknowledge

 Metaknowledge is knowledge about knowledge

and expertise.
 Most successful expert systems are restricted to
as small a domain as possible.
 In an expert system, an ontology is the
metaknowledge that describes everything known
about the problem domain.
 Wisdom is the metaknowledge of determining
the best goals of life and how to obtain them.

54
The Pyramid of Knowledge

55
 Productions

A number of knowledge-representation techniques

have been devised:
• Rules
• Semantic nets
• Frames
• Scripts
• Logic
• Conceptual graphs

56
Parse Tree of a Sentence

57
 Semantic Nets

 A classic representation technique for

propositional information
 Propositions – a form of declarative knowledge,
stating facts (true/false)
 Propositions are called “atoms” – cannot be
further subdivided.
 Semantic nets consist of nodes (objects, concepts,
situations) and arcs (relationships between them).

58
 Common Types of Links

 IS-A – relates an instance or individual to a

generic class

 A-KIND-OF – relates generic nodes to generic

nodes

59
Two Types of Nets

60
General Organization of a PROLOG
System

61
 PROLOG and Semantic Nets

 In PROLOG, predicate expressions consist of the

predicate name, followed by zero or more
arguments enclosed in parentheses, separated by
commas.
 Example:
mother(becky,heather)
means that becky is the mother of heather

62
 PROLOG Continued

 Programs consist of facts and rules in the

general form of goals.
 General form: p:- p1, p2, …, pN
p is called the rule’s head and the pi
represents the subgoals
 Example:
spouse(x,y) :- wife(x,y)
x is the spouse of y if x is the wife of y
63
 Object-Attribute-Value Triple

 One problem with semantic nets is lack of

standard definitions for link names (IS-A, AKO,
etc.).

 The OAV triplet can be used to characterize all

the knowledge in a semantic net.

64
 Problems with Semantic Nets

 To represent definitive knowledge, the link and

node names must be rigorously defined.
 A solution to this is extensible markup language
(XML) and ontologies.
 Problems also include combinatorial explosion of
searching nodes, inability to define knowledge
the way logic can, and heuristic inadequacy.

65
 Schemata

 Knowledge Structure – an ordered collection of

knowledge – not just data.
 Semantic Nets – are shallow knowledge
structures – all knowledge is contained in nodes
and links.
 Schema is a more complex knowledge structure
than a semantic net.
 In a schema, a node is like a record which may
contain data, records, and/or pointers to nodes.

66
 Frames

 One type of schema is a frame (or script – time-

ordered sequence of frames).
 Frames are useful for simulating commonsense
knowledge.
 Semantic nets provide 2-dimensional knowledge;
frames provide 3-dimensional.
 Frames represent related knowledge about
narrow subjects having much default knowledge.

67
 Frames Continued

 A frame is a group of slots and fillers that defines

a stereotypical object that is used to represent
generic / specific knowledge.
 Commonsense knowledge is knowledge that is
generally known.
 Prototypes are objects possessing all typical
characteristics of whatever is being modeled.
 Problems with frames include allowing
unrestrained alteration / cancellation of slots.

68
 Logic and Sets

 Knowledge can also be represented by symbols

of logic.

 Logic is the study of rules of exact reasoning –

inferring conclusions from premises.

 Automated reasoning – logic programming in the

context of expert systems.

69
A Car Frame

70
 Forms of Logic

 Earliest form of logic was based on the syllogism

– developed by Aristotle.
 Syllogisms – have two premises that provide
evidence to support a conclusion.
 Example:
 Premise: All cats are climbers.
 Premise: Garfield is a cat.
 Conclusion: Garfield is a climber.

71
 Venn Diagrams

 Venn diagrams can be used to represent

knowledge.
 Universal set is the topic of discussion.
 Subsets, proper subsets, intersection, union ,
contained in, and complement are all familiar
terms related to sets.
 An empty set (null set) has no elements.

72
Venn Diagrams

73
 Syllogism 三段論法

Premise: All men are

mortal
Premise: Socrates is a man
Conclusion: Socrates is mortal

Only the form is important.

Premise: All X are Y

Premise: Z is a X
Conclusion: Z is a Y

74
 Categorical Syllogism

 Syllogism: a valid deductive

argument having two premises and a
conclusion.

major premise: All M are P

minor premise: All S is M
Conclusion: All S is P

M middle term
P major term
S minor term
75
 Categorical Statements
Form Schema Meaning

A All S is P universal affirmative

E No S is P universal negative

I Some S is P particular affirmative

O Some S is not P particular negative

76
 Figure 1 2 3 4
Major Premise M P PM MP PM
Minor Premise S M SM MS MS

Mood
AAA-1 EAE-1 IAI-4
All M is P No M is P Some P is M
All S is M All S is M All M
is S
All S is P No S is P Some S is P

77
 Propositional Logic

 Formal logic is concerned with syntax of

statements, not semantics.
 Syllogism:
• All goons are loons.
• Zadok is a goon.
• Zadok is a loon.

 The words may be nonsense, but the form is

correct – this is a “valid argument.”

78
Intersecting Sets

79
 Boolean Logic

 Defines a set of axioms consisting of symbols to

represent objects / classes.
 Defines a set of algebraic expressions to
manipulate those symbols.
 Using axioms, theorems can be constructed.
 A theorem can be proved by showing how it is
derived from a set of axioms.

80
 Features of Propositional Logic

 Concerned with the subset of declarative

sentences that can be classified as true or false.
 We call these sentences “statements” or
“propositions”.
 Paradoxes – statements that cannot be classified
as true or false.
 Open sentences – statements that cannot be
answered absolutely.

81
 Features Continued

 Compound statements – formed by using logical

connectives (e.g., AND, OR, NOT, conditional,
and biconditional) on individual statements.

 Material implication – p  q states that if p is

true, it must follow that q is true.

 Biconditional – p  q states that p implies q and

q implies p.

82
 Features Continued

 Tautology – a statement that is true for all

possible cases.

 Contradiction – a statement that is false for all

possible cases.

 Contingent statement – a statement that is neither

a tautology nor a contradiction.

83
Truth Tables

84
 Rule of Inference
 Modus ponens, way assert
direct reasoning, law of detachment
assuming the antecedent

 Modus tollens, way deny

indirect reasoning, law of
contraposition
assuming the antecedent

85
 Modus Ponens
If there is power, the computer will
work
There is power
------------------------------
The computer will work

AB p, p ->q; q
A
--------------
B 86
 Modus tollens

pq conditional p -> q

~q converse q -> p
-------------- inverse ~p ->
~q
~p contrapositive
~q -> ~p

87
 Formal Logic Proof
Chip prices rise only if the yen rises.
The yen rises only if the dollar falls and
if the dollar falls then the yen rises.
Since chip prices have risen,
the dollar must have fallen.

CY C = chip prices rise

(Y  D) ^ (D  Y) Y = yen rises
C D = dollar falls
-------------------------
D

88
 Formal Logic Proof
CY
(Y  D) ^ (D  Y)
C
D

1. C  Y premise
2. (Y  D) ^ (D  Y) premise
3. C
4. Y == D 2 Equivalence
5. CD 1 Substitution
6. D 3,5 modus ponens

89
 Resolution

Normal form
 Conjunctive normal form
(P1vP2v…)^(Q1vQ2…)^…
(Z1vZ2…)
 Kowalski clausal form
A1, A2, …., An  B1, B2, ….,
Bm
 Horn clause
A1, A2, …., An  B
90
 Method of Contradiction

Av B
(AvB)^(Av~B)
A v ~B A v (B^~B)
-------------- A
A

91
 Forward Reasoning

T: A  B
BC 
CONCLUSION?
CD
A

Resolve by modus
ponens

92
 Backward Reasoning

What if T is very large?

T may support all kinds of
inferences
which have nothing to do
with
the proof of our goal
Combinational explosion

 Use Backward Reasoning

93
 Resolution Refutation

 To refute something is to prove it

false
 Refutation complete:
Resolution refutation will
terminate in
a finite steps if there is a
contradiction
 Example: Given the argument
AB
BC
CD
-------------
AD 94
 Example

AB
BC
CD
-------------
AD

To prove that A  D is a theorem by resolution refutation:

1. A  D equ ~A v D convert to disjunction form

2. ~(~A v D) equ A ^ ~D negate the conclusion
3. A  B equ ~A v B
B  C equ ~B v C
C  D equ ~C v D
4. (~A v B) ^ (~B v C ) ^ (~C v D) ^ A ^ ~D
resolution
=> nil false
95
 Method of Contradiction

Av B
(AvB)^(Av~B)
A v ~B A v (B^~B)
-------------- A
A

96
 Propositional Logic
 Symbolic logic for manipulating
proposition
 Proposition, Statement, Close
sentence:
a sentence whose truth value can
be
determined.
 Open Sentence: a sentence which
contains variables
 Combinational explosion
 97
 Predicate Logic
 Predicates with arguments
on-top-of(A, B)
 Variables and Quantifiers
Universal (x)(Rational(x)  Real(x))
Existential (x)(Prime(x))
 Functions of Variables
(x)(Satellite(x))  (y)(closest(y,
earth)^on(y,x))

(x)(man(x)  mortal(x))
^ man(Socrates)
=> mortal(Socrates)

98
 Universal Quantifier

 The universal quantifier, represented by the

symbol  means “for every” or “for all”.
(x) (x is a rectangle  x has four sides)
 The existential quantifier, represented by the
symbol  means “there exists”.
(x) (x – 3 = 5)
 Limitations of predicate logic – most quantifier.

99
 First Order Predicate Logic
 Quantification not over predicate or
function symbols
 No MOST quantifier, (counting required)
 Can not express things that are
sometime true
=> Fuzzy Logic

100
 Syllogism in Predicate Logic
Type Scheme Predicate
Representation
A All S is P (x)(S(x) -> P(x))
E No S is P (x)(S(x) ->
~P(x))
I Some S is P (x)(S(x) ^ P(x))
A Some S is not P (x)(S(x) ^
~P(x))

101
 Rule of Universal Instantiation

(x)p(x) => p(a) p: any

proposition or
propositional
function
a: an instance

102
 Formal Proof

(x)(H(x)->M(x)) All men are mortal

H(s) Socrates is a man
Þ M(s) => Socrates is mortal

1. (x)(H(x)->M(x)) premise
2. H(s)
3. H(s)->M(s) universal
instantiation
4. M(s) 2,3 modus ponens

103
 Well-formed Formula

1. An atom is a formula
2. If F and G are formula, then ~(F),
(FvG),
(F^G),(F->G), and (F<->G) are
formula
3. If F is a formula, and x is a free
variable in F, then (x)F and
(x)F are formula
4. Formula are generated only by a
finite number of applications of 104
 Summary

 We have discussed:
 Elements of knowledge
 Knowledge representation
 Some methods of representing knowledge
 Fallacies may result from confusion between
form of knowledge and semantics.
 It is necessary to specify formal rules for expert
systems to be able to reach valid conclusions.
 Different problems require different tools.

105