Introduction to

Expert
Systems

• Introduction
• Main components
• Inference mechanisms

(C) Copyright 2022 by Prof.Abeer El-Korany 1

 Knowledge Based Systems
The goal is to facilitate intelligent interaction
with the user based on:
– the identification of the appropriate information
– the effective utilization of the appropriate information
– the control of the appropriate information
in order to fulfill specific user goals

(C) Copyright 2022 by Prof.Abeer El-Korany 2

 Main Components of an KBS

Expertise

Knowledge Base

Facts / Information

Inference Engine
Expertise

(C) Copyright 2022 by Prof.Abeer El-Korany 3

 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

(C) Copyright 2022 by Prof.Abeer El-Korany 4

 Basic Functions
of Expert Systems

(C) Copyright 2022 by Prof.Abeer El-Korany 5

 Tasks of Expert Systems

(C) Copyright 2022 by Prof.Abeer El-Korany 6

 Development of an Expert System

(C) Copyright 2022 by Prof.Abeer El-Korany 7

 ES Tools
• ES languages
– higher-level languages specifically designed for
knowledge representation and reasoning
• ES shells
– an ES development tool/environment where the
user provides the knowledge base
– CLIPS, JESS, Mycin, Babylon, G2, ...

8 Copyright
(C)
2022 by
 Structure of a Rule-Based System

(C) Copyright 2022 by Prof.Abeer El-Korany 9

 Rule-Based System
• Rules can be used to formulate a theory of human
information processing (Newell & Simon)
– rules are stored in long-term memory (KB)
– temporary knowledge is kept in short-term memory
(agenda)
– Input or thinking triggers the activation of rules
– activated rules may trigger further activation
– a cognitive processor combines evidence from
currently active rules
• This model is the basis for the design of many
rule-based systems
– also called production systems

(C) Copyright 2022 by Prof.Abeer El-Korany

10
◼ The human mental process is internal, and it is too
complex to be represented as an algorithm.
However, most experts are capable of expressing
their knowledge in the form of rules for problem
solving.

IF the ‘traffic light’ is green

THEN the action is go

IF the ‘traffic light’ is red

THEN the action is stop

(C) Copyright 2022 by Prof.Abeer El-Korany 11

 Example Rules
IF … THEN Rules
antecedent
Rule: Red_Light
(left-hand-side)
IF the light is red
THEN stop
Rule: Green_Light consequent
IF the light is green (right-hand-side)
THEN go
antecedent (left-hand-side)
Production Rules
the light is red ==> stop
consequent
(right-hand-side)
the light is green ==> go
(C) Copyright 2022 by Prof.Abeer12
El-Korany
 Example: Rule-Base to determine the ‘grade’:

1. study good_grade
2. not_study bad_grade
3. sun_shines go_out
4. go_out not_study
5. stay_home study
6. bad_weather stay_home

Q1: If the weather is bad, do you get a good or bad

grade?

(C) Copyright 2022 by Prof.Abeer El- 13

Korany
 Rule-Based Systems
• knowledge is encoded as IF … THEN rules
– these rules can also be written as production rules
• the inference engine determines which rule
antecedents are satisfied
– the left-hand side must “match” a fact in the working
memory
• satisfied rules are placed on the agenda
• rules on the agenda can be activated (“fired”)
– an activated rule may generate new facts through its right-
hand side
– the activation of one rule may subsequently cause the
activation of other rules
(C) Copyright 2022 by Prof.Abeer El-Korany 14
 Rule-based reasoning: rules
• The essence of a rule-based reasoning system is that it goes
through a series of cycles.

• In each cycle, it attempts to pick an appropriate rule from its

collection of rules, depending on the present circumstances, and to
use it as described above.

• Because using a rule produces new information, it's possible for

each new cycle to take the reasoning process further than the cycle
before. This is rather like a human following a chain of ideas in
order to come to a conclusion.

(C) Copyright 2022 by Prof.Abeer El-Korany

15
 Reasoning with production rules

INFERENCE
WORKING ENGINE
MEMORY

PATTERN EXECUTION
MATCHER ENGINE
RULE
BASE
AGENDA

(C) Copyright 2022 by Prof.Abeer El-Korany 16

Reasoning with production rules(cont.)

Architecture of a typical production

system:
observed data

working memory
select modify

Knowledge interpreter
base fire output

(C) Copyright 2022 by Prof.Abeer El-Korany

17
Reasoning with production rules(cont.)
Architecture of a typical production
system:

New information

working memory
select modify

Knowledge interpreter
base fire output

18
(C) Copyright 2022 by Prof.Abeer El-Korany
Reasoning with production rules(cont.)
Architecture of a typical production
system:

New information

select working memory

modify

Knowledge interpreter
base fire output

19
(C) Copyright 2022 by Prof.Abeer El-Korany
Reasoning with production rules(cont.)
Architecture of a typical production
system:

New information

working memory
select
modify

Interpreter
Knowledge executes
base fire actions
output

20
(C) Copyright 2022 by Prof.Abeer El-Korany
Reasoning with production rules(cont.)

Architecture of a typical production

system:
New information

working memory
select modify

Interpreter
Knowledge executes
base fire actions output

21
(C) Copyright 2022 by Prof.Abeer El-Korany
 Inference Engine Cycle
• Describes the execution of rules by the inference engine
– match
• Find rules whose antecedents are satisfied
• Add them to agenda
– conflict resolution
• select the rule with the highest priority from the agenda
– execution
• perform the actions on the consequent of the selected rule
• Update the agenda (remove the fired rule from the agenda and add others)

• The cycle ends when no more rules are on the agenda, or when
an explicit stop command is encountered

(C) Copyright 2022 by Prof.Abeer El-Korany

22
 Chaining

• Chain – a group of multiple inferences that

connect a problem with its solution
• A chain that is searched / traversed from a
problem to its solution is called a forward chain.
• A chain traversed from a hypothesis back to the
facts that support the hypothesis is a backward
chain.
• Problem with backward chaining is find a chain
linking the evidence to the hypothesis.
(C) Copyright 2022 by Prof.Abeer El-
Korany 23
 Control schemes
Two kinds of control in rule-based systems:
– Forward chaining
– Backward chaining.

• Forward chaining starts with the facts, and sees

what rules apply (and hence what should be done)
given the facts.

• Backward chaining (much like Prolog) starts with

something to find out, and looks for rules that will
help in answering it.

(C) Copyright 2022 by Prof.Abeer El-Korany24

 Forward and Backward Chaining
• Forward chaining (data-driven)
– Forward chaining is bottom-up reasoning, i.e. reasoning
from facts to goals. Reasoning (moving) from facts to the
conclusion
– As soon as facts are available, they are used to match
antecedents of rules
– This process of cascading triggering of rules is called
‘chaining’ as a chain of rules may be fired.
• Backward chaining (goal/query-driven)
– Starting from a hypothesis (query-goal, supporting rules
and facts are sought until all parts of the antecedent of the
hypothesis are satisfied)
– often used in diagnostic and consultation systems

25 Copyright
(C)
2022 by
 Forward chaining Example
• Consider the simple fire example
R1: IF hot AND smoky THEN there is a fire
R2: IF alarm_beeps THEN smoky
R3 IF there is a fire THEN switch_on_sprinklers

• Working memory initially contains two facts:

Working memory
alarm_beeps .1
hot .2

Follow the algorithm: First cycle. Working memory

Find all rules with satisfied conditions : R2 alarm_beeps .1
Choose one: R2 Hot .2
Perform actions: smoky. smoky .3

26 Copyright
(C)
2022 by
 Forward chaining Example (cont.)
• Second cycle:
– Find all rules with conditions satisfied : R1
– Choose one and apply action: there is a fire
. Working memory

alarm_beeps .1
Hot .2
Third cycle Smoky .3
Rules with conditions satisfied:there
R3 is a fire. .4
apply action: switch_on_sprinklers

27 Copyright
(C)
2022 by
 Forward chaining
• Facts are held in a working memory

• A rule can be activated if all parts of the antecedent are satisfied

• Matches the antecedents of available rules with the current state until it
finds a rule all of whose antecedents are satisfied

• If the system supports variables, then matching may require unification

• The inference engine can infer further facts. These new facts may in turn
trigger further rules, which may generate further facts.

• Then repeats the cycle

• A rule can be activated if all parts of the antecedent are satisfied

• This defines a forward-chaining inference procedure because it moves

“forward” from the KB to the goal

28 Copyright
(C)
2022 by
 Forward Chaining Inferencing
(cont.)
1. Scan the rules looking for ones whose
premises match the contents of the working
memory.
2. Fire the rule which was found.
3. Place its conclusion in the working memory.
4. Until no additional rule fire, go to 1.

29 Copyright
(C)
2022 by
 Inference Network Representation
• We can represent forward chaining with rules as a network:

has hair
is a mammal

eats meat
is a carnivore

tawny color
is a tiger
black stripes
has a tail

(C) Copyright 2022 by Prof.Abeer El-Korany 30

 Backward Chaining
• Backward chaining is top-down reasoning, i.e. reasoning from
goals to facts.

• Basic algorithm, to prove goal G:

– If G is in the initial facts, it is proven.
– Otherwise, find a rule which can be used to conclude G, and try to prove
each of that rule’s conditions.
• Avoid repeated work: check if new subgoal
– Has already been proved true
– Has already failed

• This allows rather more focused style of reasoning. (Forward

chaining may result in a lot of irrelevant conclusions added to
working memory.)
31 Copyright
(C)
2022 by
 Backward Chaining Example
• Consider the simple fire example
• Following the algorithm: First cycle.
– Determine the goal Working memory
– Set R3 switch on the sprinklers as G1 alarm_beeps .1
hot .2
• G1: switch_on_sprinklers
– Is it in initial facts? No.
– Is there a rule which adds this as a conclusion? Yes, R3
– Set condition of R3 as new goal

• G2: there is fire.

• Second cycle:
– Is it in initial facts? No.
– Is there a rule which adds this as a conclusion? Yes, R1

32 Copyright
(C)
2022 by
 Backward Chaining Example (cont.)
• Set conditions as new goals: G3: hot, G4: smoky.
• G3: hot, G4: smoky
• Third cycle:
– Try to prove G3: hot. In initial facts.
– Try to prove G4: smoky. Conclusion of rule R2

• Fourth cycle:
• G5: alarm_beeps.
– In initial facts, so all done…
– Proved hypothesis switch_on_sprinklers.

33 Copyright
(C)
2022 by
 Forward &Backward Chaining
Example (Cont.)
• A forward chaining system would note first that it
knew A and B and could therefore conclude C–
then that it knew B and C and could then conclude
D– finally note that fortuitously, D was what it
wanted to prove

• A backward chaining system would– note first

that it needed to prove C and B to prove D– then
that it already knew B then that it needed to prove
A and B to prove C finally that it already knew A
and B.

34 Copyright
(C)
2022 by
 • Example
An expert system used to make classification between Ostrich and
Duck have the following rules: -

R1: IF have wing = yes and mammel = yes THEN bird = yes
R2: IF bird = yes and neck length = long THEN can-fly = no
R3: IF bird = yes and leg length = small THEN can-fly = no
R4: IF can-fly = no and leg length = small and neck length = small
THEN Duck = yes , Ostrich = no
R5: IF can-fly = no and leg length = long and neck length = long
THEN Duck = no , Ostrich = yes

Trace the algorithm of reasoning with the following user inputs:

[mammel = yes , have wing = yes , leg length = small ,neck length = small]

(C) Copyright 2022 by Prof.Abeer El-Korany 35

 Apply backward chaining: Goals (Duck Ostrich).
Select the first top goal Duck .
– Rules R4 & R5 that derive that goal placed on the stack .
– Rule R4 examined , the first parameter (can fly =fly) examined to see if it
is in the database (which is not) So that makes can fly as sub-goal (and
go back to step 2).
– R2 & R3 derive value for (can fly =fly)
– Rule R2 examined , examine-> R1 derive value for (bird)
– Rule R1 examined ,
– Premises of R1 statisfied and get (bird=yes)
– Rule R2 examined ->R2 fails
– Rule R3 examined ->R3 satisfied (fly=no)
– Rule R4 all three premises true then it is satisfied and drive (Duck=yes ,
Ostrich=no)

(C) Copyright 2022 by Prof.Abeer El-Korany 36

 Forward Chaining Example
Rules:
R1: IF on-cl(green) F1- on-cl(green)
THEN put-on-cl(produce)
F2- weighs =15 lbs
R2: IF on-cl(packed in small containers)
THEN put-on-cl(delicacy)
R3: IF on-cl(refrigerated )OR put-on-cl(produce)
THEN put-on-cl(perishable)
R4: IF put-on-cl(perishable)& (weighs=15 lbs) & (inexpensive)
THEN put-on-cl(staple)
R5: IF on-cl(meat) & put-on-cl(perishable) & (weighs=15 lbs)
THEN put-on-cl(turkey)
R6: IF (weighs=15 lbs) &put-on-cl(produce)
THEN put-on-cl(watermelon)

(C) Copyright 2022 by Prof.Abeer El-Korany

 Example - Execution
Fact List Applicable Rules
R1
(green, weighs =15 lbs)

(C) Copyright 2022 by Prof.Abeer El-Korany 38

 Example - Execution
Fact List Applicable Rules
(green, weighs=15 lbs) R1
(green, weighs=15 lbs, R1, R3, R6
put-on-cl(produce)

(C) Copyright 2022 by Prof.Abeer El-Korany 39

 Example - Execution
Fact List Applicable Rules
(green, weighs =15 lbs)
(green, weighs =15 lbs, R1, R3, R6
put-on-cl(produce) )

(green, weighs =15 lbs, R3, R6

put-on-cl(produce), perishable)

(C) Copyright 2022 by Prof.Abeer El-Korany 40

 Example - Execution
Fact List Applicable Rules
R1
(green, weighs = 15 lbs)
(green, weighs =15 lbs, R1, R3, R6
produce)

(green, weighs =15 lbs, R1, R3, R6

produce, perishable)

(green, weighs =15 lbs, R1, R3, R6

produce, perishable,
watermelon) ALL DONE!
(C) Copyright 2022 by Prof.Abeer El-Korany 41
Example:Auto Diagnosis

engine-gets-gas Rule 1: if the engine is getting gas,

engine-turns and the engine will turn over,
Goal (problem(spark-plugs)) then the problem is spark plugs.

not(engine-turns) Rule 2: if the engine does not turn over,

not(lights-on) and the lights do not come on,
then (problem (battery or cables)) . Goal(problem(battery-or-cables))

not(engine-turns) Rule 3: if the engine does not turn over,

lights-on and the lights do come on,
Goal(problem(starter-motor)) then the problem is the starter motor.

gas-in-tank Rule 4: if there is gas in the fuel tank,

gas-in-carburettor and there is gas in the carburettor,
engine-gets-gas then the engine is getting gas.

(C) Copyright 2022 by Prof.Abeer El-Korany

42
 Matching Rules Working Memory
there is gas in the fuel tank
R4 there is gas in the carburettor
the engine will turn over

Reasoning tree
starter-motor battery-or-cables spark-plugs

not
not
not

lights-on engine will turn over Engine is getting gas

gas-in-feul tank gas-in-carburettor

(C) Copyright 2022 by Prof.Abeer El-Korany 43

 Simple Medical Expert System
• Rules
1. If headache then prescribe pain killer
2. If headache and sore throat and coughing then
diagnose flu
3. If tired and headache then diagnose glandular
fever
4. If tired and sore throat then diagnose
tonsillitis
5. If tired than diagnose stress

44 Copyright
(C)
2022 by
 Simple Medical Expert System
• User interaction
– Patient: I have a headache
– System: I suggest you take a pain killer. Do you have a sore
throat?
– Patient: Yes, I do.
– System: Have you been coughing?
– Patient: No,
– System: Have you been feeling tired recently?
– Patient: Yes, I have.
– System: It sounds as though you may have glandular fever or
tonsillitis.

45 Copyright
(C)
2022 by
 Forward vs. Backward Chaining
Forward Chaining Backward Chaining

planning, control diagnosis

data-driven goal-driven (hypothesis)
bottom-up reasoning top-down reasoning
find possible conclusions find facts that support a given
supported by given facts hypothesis
similar to breadth-first search similar to depth-first search
antecedents (LHS) control consequents (RHS) control
evaluation evaluation
(C) Copyright 2022 by
46
Prof.Abeer El-Korany 46
 Advantages of Rule Systems
• Separation of control (inference) from knowledge
• Modularity of knowledge (A rule is an independent chunk
of knowledge)
• Ease of expansion ( Proportional growth of intelligence)
• Derivation of explanations from rigid syntax
• Consistency checking
• Utilization of uncertain knowledge
– IF it looks like rain THEN I should probably carry an umbrella
– IF Weather looks like rain THEN Carry an umbrella CF 80
• Can incorporate variables
IF ?Student marks is adequate
THEN ?Student can graduate

(C) Copyright 2022 by Prof.Abeer 47

El-Korany