Introduction to Artificial Intelligence
Semantic Network
By: 1
Boreshban [email protected]
2
Knowledge Representation
• Logic isn‟t the only method of representing
knowledge.
• There are other methods which are less general, but
more natural, and arguably easier to work with:
▫ Semantic Nets
▫ Conceptual Dependency
▫ Frames
▫ Scripts
3
Semantic networks eat does
Collins & Quillian
(1969) skin has
animal
fly wings can’t
feathers can has ostrich
has
isa
isa bird
isa
seagull isa isa
does
robin canary sing
4
Semantic Net
• Basic Constructs
▫ Node-Object, Concept
▫ Links-Relation
Tweety isa Robin isa Bird
has-part
Wings Wings Wings
property inheritance
-Property Inheritance is the main inference
mechanism!
5
Nodes and Arcs
• Arcs define binary relations which hold between
objects denoted by the nodes.
mother age
Sue John 5
mother (john, sue)
age (john, 5)
father wife (sue, max)
age
age (max, 34)
…
34 Max
6
Semantic Net (Cont.)
Example
Sold-by Furniture
Furniture store
isa
Chair has-part
Seat
isa
Park My-Chair
owner
isa cover
Person Leather Black
cover
7
Non-binary relations
• We can represent the generic give event as a relation
involving three things:
▫ A giver
▫ A recipient
▫ An object recipient giver
Mary GIVE John
object
book
8
More examples
• “John is taller than Bill.”
John Bill
Is-taller
John Bill
height height
H1 Number H2
isa isa
greater-than
9
Network Features
• In “abstract” graphs, the graph is simpler, but
processing capability on graph is low.
• On the other hand, with simple relations, the graph
is more complicated, but processing capability is
high.
• We prefer the second situation!
10
Reasoning with Semantic Nets
• Matching
isa
Tweety Robin Bird
isa isa
Tweety ?
isa
• What is Tweety?
11
Advantages of Semantic nets
• Easy to visualize
• Related knowledge is easily clustered.
• Efficient in space requirements
▫ Objects represented only once
12
Problems of Semantic Net
1. Different people use different nets to represent the
same thing.
John Mary
married
Event
isa
Marriage
isa
M1
male female
John Mary
13
Problems of Semantic Net
2. Same Net interpreted differently by different person.
Jack Tom
Father-of
3. Quantification and intentional concepts are hard to
represent.
▫ Some birds fly
▫ All the birds sing some of the songs
▫ Some of the birds sing all the songs
14
Dealing with Exceptions
Fly
CAN
CANNOT
Bird
ISA
Ostrich
ISA
Henry
15
Dealing with Exceptions(Continue)
Animal
ISA
Travel ISA
Fly By Bird Penguin
ISA Travel
By
Eagle
Walk
16
Some Issues
• Use simple concepts
• Try to construct complex concepts with these simple
concepts
• For example in kinship relation
▫ Use simple concepts such as
Brother, Sister, Mother, Father, Daughter, Son,
Husband
▫ Construct complex concepts (relations) such as
Cousin, Niece, Nephew , Wife, Grandfather, …
with former simple concepts
17
First Order Login in Semantic Net.
• Every dog has bitten a postman.
18
dog bitten postman
ISA ISA ISA
Agent Victim
d b p
form
ISA
GS g
19
The graph meaning
• “GS” = “Generalized Statement” and show all general
concepts in the world!
• In the graph, we generalize „d‟ with „g‟.
• The dashed box shows the domain of the quantifier,
and „form‟ shows relation between quantifier and box.
20
Example2 for FOL
• There exists a postman that Every dog has bitten
him.
• Here, you only must exclude the „p‟ from dashed
box!
21
dog bitten postman
ISA ISA ISA
Agent Victim
d b p
form
ISA
GS g
