Вы находитесь на странице: 1из 39

Artificial Intelligence

Representing Commonsense
Knowledge
L. Manevitz

All rights reserved © L. Manevitz Lecture 7 1


Definitions
• Representation – a set of syntactic and semantic
conventions that make it possible to describe
things.
• Syntax – specifies the symbols that may be used
and the ways those symbols may be arranged.
• Semantics – specifies how meaning is embodied
in the symbol arrangements allowed by the
syntax.

All rights reserved © L. Manevitz Lecture 7 2


Semantic Network Approach
• Nodes and Slots:
Nodes are objects,
or classes,
or properties.
Slots are of different types.

All rights reserved © L. Manevitz Lecture 7 3


A Semantic Network

Mammal

Is-a
has-part
Person Nose

uniform- instance
color team
Blue Pee-Wee-Reese Brooklyn-Dodgers

All rights reserved © L. Manevitz Lecture 7 4


Representing Nonbinary
Predicates
• Unary Predicates can be rewritten as
binary ones.

man(Marcus)
could be rewritten as
instance(Marcus,Man)

All rights reserved © L. Manevitz Lecture 7 5


Representing Nonbinary
Predicates cont.
• N-Place Predicates
score(Cubs,Dodgers,5-3)
becomes Game
visiting- Is-a
team score
Cubs G23 5-3

home-team
Dodgers
All rights reserved © L. Manevitz Lecture 7 6
A Semantic Net Representing
a Sentence
“John gave the book to Mary.”
Give Book

instance instance
agent object
John EV7 BK23

beneficiary

Mary

All rights reserved © L. Manevitz Lecture 7 7


Some Important Distinctions
First try:
height
John 72

Second try:
John Bill

height height
greater-than
H1 H2
All rights reserved © L. Manevitz Lecture 7 8
Some Important Distinctions
cont.
Third try:
John Bill

height height
greater-than
H1 H2
value

72

All rights reserved © L. Manevitz Lecture 7 9


Partitioned Semantic Nets
a) The dog bit the mail carrier.

Dogs Bite Mail-carrier

Is-a Is-a Is-a


assailant victim
d b m

All rights reserved © L. Manevitz Lecture 7 10


Partitioned Semantic Nets cont.
a) Every dog has bitten a mail carrier.
SA

GS Dogs Bite Mail-carrier

Is-a S1 Is-a Is-a Is-a


form assailant victim
d b m
g

All rights reserved © L. Manevitz Lecture 7 11


Partitioned Semantic Nets cont.
a) Every dog in town has bitten the
constable.
SA
Dogs

GS Bite Constables
Town-Dogs

Is-a S1 Is-a Is-a Is-a


form assailant victim
d b c
g
All rights reserved © L. Manevitz Lecture 7 12
Partitioned Semantic Nets cont.
a) Every dog has bitten every mail carrier.
SA
Dogs Bite Mail-carrier

S1 Is-a Is-a Is-a


assailant victim
d b m

form

GS Is-a g
All rights reserved © L. Manevitz Lecture 7 13
Inheritance
• Is-a slot – appears between objects and
classes.

• ako slot – appears between subsets.

All rights reserved © L. Manevitz Lecture 7 14


Inheritance -Procedure
F the given node; S the given slot;
2. Form a Queue of F and all class nodes connected to F
via Is-A node. F is at top of Queue.
3. Until Queue is empty or default has been found
determine if the first element of Queue has a value in
its S slot:
a. Yes – a value has been found.
b. No – remove the first element from Queue and add the nodes
related to the first element by AKO slots to the end of Queue.
4. If a value has been found say that this is the default
value of F’s S slot.
Otherwise announce Failure.
All rights reserved © L. Manevitz Lecture 7 15
Inheritance - Example

Block

ako ako
shape shape
Wedge Triangular Brick rectangular

Is-a Is-a
Wedge18 Brick12

All rights reserved © L. Manevitz Lecture 7 16


If-needed Inheritance -Procedure
F the given node; S the given slot;
2. Form a Queue of F and all class nodes connected to F
via Is-A node. F is at top of Queue.
3. Until Queue is empty or successful if-needed procedure
has been found determine if the first element of Queue
has a procedure in the If-Needed facet of its S slot:
a. Yes – if the procedure produces a value than a value has been
found.
b. No – remove the first element from Queue and add the nodes
related to the first element by AKO slots to the end of Queue.
4. If a value has been found say that the value found is the
value of F’s S slot.
Otherwise announce Failure.
All rights reserved © L. Manevitz Lecture 7 17
If-needed Inheritance - Example

Weight (if-needed)
Block-weight-
Block procedure

Brick
Volume 400
Brick12
Density 11

All rights reserved © L. Manevitz Lecture 7 18


Example cont.
Weight is activated
by request for the
Block weight of Brick12 !

Brick
Weight 4400

Volume
Brick12 400

Density 11
All rights reserved © L. Manevitz Lecture 7 19
Default Inheritance Procedure
F the given node; S the given slot;
2. Form a Queue of F and all class nodes connected to F
via Is-A node. F is at top of Queue.
3. Until Queue is empty or default has been found
determine if the first element of Queue has a value in
the Default facet of its S slot:
a. Yes – if the first element has a value than a value has been
found.
b. No – remove the first element from Queue and add the nodes
related to the first element by AKO slots to the end of Queue.
4. If a value has been found say that the value found is the
default value of F’s S slot.
Otherwise announce Failure.
All rights reserved © L. Manevitz Lecture 7 20
Default Inheritance - Example
Has no default
color therefore
probably Blue Color (Default)
because of Block Blue
Block’s default
color ! ako ako
Color (Default)
Wedge Brick Red

Is-a Is-a
Wedge18 Brick12
All rights reserved © L. Manevitz Lecture 7 21
Perspective -Example
rectangular
Play Commemorate
shape Purpose
Purpose
Brick Structure Toy Gift

Is-a Is-a Is-a


Is-a
Structure Toy Gift
Brick12
perspective perspective perspective

Purpose
Support

All rights reserved © L. Manevitz Lecture 7 22


Special Links - Summary
• IS-A and AKO links make class
membership and subclass-class relations
explicit, facilitating the movement of
knowledge from one level to another.
• VALUE facets make values explicit.

All rights reserved © L. Manevitz Lecture 7 23


Special Links – Summary cont.
• IF-NEEDED facets make procedures purposes
explicit, and they relate procedures to the classes
those procedures are relevant to.
• DEFAULT facets make likely values explicit
without implying certainty.
• Perspectives make context sensitivity explicit,
preventing confusion and ambiguity.

All rights reserved © L. Manevitz Lecture 7 24


Frames
• Frames : A collection of nodes that
describe a stereotyped object, act or event.

• Example : newspaper report.

All rights reserved © L. Manevitz Lecture 7 25


Earthquake Example
Disaster-event Event
Killed Place
Injured Day
Homeless Time
Damage
Earthquake Social-event
Magnitude
Number-of-
Fault guests
Flood Host
Crest
River Birthday-party
Hurricane Birthday-
Wind-speed person
Name Age
All rights reserved © L. Manevitz Lecture 7 26
Earthquake Example cont.
Earthquake Hits Lower Slabovia
• Today an extremely serious earthquake of
magnitude 8.5 hit Lower Slabovia killing
25 people and causing $500,000,000 in
damage. The president of Lower Slabovia
said the hard-hit area near the Sadie
Hawkins fault has been a danger zone for
years.
All rights reserved © L. Manevitz Lecture 7 27
Earthquake Example cont.
Earthquake13
place Lower Slabovia

day Today

fatalities
25
damage
500,000,000
magnitude
8.5

fault Sadie Hawkins


All rights reserved © L. Manevitz Lecture 7 28
Earthquake Summary Pattern
• An earthquake occurred in value in
location slot value in day slot. There were
value in fatalities slot fatalities and value
in damage slot in property damage. The
magnitude was value in magnitude slot on
the Richter scale, and the fault involved
was the value in fault slot.

All rights reserved © L. Manevitz Lecture 7 29


Instantiated Earthquake
Summary Pattern
• An earthquake occurred in Lower Slabovia
today . There were 25 fatalities and $500
million in property damage. The magnitude
was 8.5 on the Richter scale, and the fault
involved was the Sadie Hawkins.

All rights reserved © L. Manevitz Lecture 7 30


Earthquake Example cont.
Earthquake Study Stopped
Today, the President of Lower Slabovia killed 25
proposals totaling $500 million for research in
earthquake prediction. Our Lower Slabovian
correspondent calculates that 8.5 research
proposals are rejected for every one approved.
There are rumors that the President’s science
advisor, Sadie Hawkins, is at fault.

All rights reserved © L. Manevitz Lecture 7 31


Earthquake Example cont.
• The Earthquake Study Stopped story
could be summarized, naively, as though it
were the story about an actual earthquake,
producing the same frame as the
Earthquake Hits Lower Slabovia story
does.

All rights reserved © L. Manevitz Lecture 7 32


Scripts

All rights reserved © L. Manevitz Lecture 7 33


Scripts
• Example - Restaurant script.
Script: Restaurant Roles: S=Customer
Track: Coffee Shop W=Waiter
Props: Tables C=Cook
Menu M=Cashier
F=Food O=Owner
Check
Money
All rights reserved © L. Manevitz Lecture 7 34
Restaurant Example cont.
Entry conditions : S is hungry
S has money

Results : S has less money


O has more money
S is not hungry
S is pleased (optional)
All rights reserved © L. Manevitz Lecture 7 35
Restaurant Example cont.
Scene 1: Entering

S PTRANS S into restaurant


S ATTEND eyes to tables
S MBUILD where to sit
S PTRANS S to table
S MOVE S to sitting position
All rights reserved © L. Manevitz Lecture 7 36
Restaurant Example cont.
Scene 2: Ordering
(menu on table) (W brings menu) (S asks for menu)
S PTRANS menu to S S MTRANS signal to W
W PTRANS W to table
S MTRANS ‘need menu’ to W
W PTRANS W to menu
S MTRANS W to table
*S MBUILD choice of F W PTRANS W to table
S MTRANS signal to W W ATRANS menu to S
W PTRANS W to table
S MTRANS ‘I want F’ to W
W PTRANS W to C
W MTRANS (ATRANS) to C
C MTRANS ‘no F’ to W
W PTRANS W to S
W MTRANS ‘no F’ to S C DO (prepare F script) to
(go back to *) or Scene 3
(go to Scene 4 at no pay
All rights reserved © L. Manevitz Lecture 7 37
path)
Restaurant Example cont.
Scene 3 : Eating

C ATRANS F to W
W ATRANS F to S
S INGEST F
(Option : Return to Scene 2 to order more;
otherwise go to Scene 4)

All rights reserved © L. Manevitz Lecture 7 38


Restaurant Example cont.
Scene 4 : Exiting

S MTRANS to W
(W ATRANS check to S)
W MOVE (write check)
W PTRANS W to S
W ATRANS check to S
S ATRANS tip to W
S PTRANS S to M
S ATRANS money to M
S PTRANS S to out of restaurant
(No pay path)

All rights reserved © L. Manevitz Lecture 7 39

Вам также может понравиться