Artificial Intelligence and Knowledge Representation

boorishadamantΤεχνίτη Νοημοσύνη και Ρομποτική

29 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

124 εμφανίσεις

Artificial Intelligence and
Knowledge Representation

WHAT MAKES THE COMPUTER
INTELLIGENT?



Speed

of

computation


Filters

out

and

displays

only

meaningful

responses

or

solutions

to

a

specific

question


Algorithms

splits

task

into

subtasks



recursion


Neural

networks
.

WHY ARTIFICIAL INTELLIGENCE



Unlike

humans,

computers

have

trouble

understanding

specific

situations,

and

adapting

to

new

situations
.



Artificial

Intelligence

improves

machine

behavior

in

tackling

such

complex

tasks,

based

on

abstract

thought,

high
-
level

deliberative

reasoning

and

pattern

recognition
.


Artificial

Intelligence

can

help

us

understand

this

process

by

recreating

it,

then

potentially

enabling

us

to

enhance

it

beyond

our

current

capabilities


KNOWLEDGE REPRESENTATION?


EXAMPLE
:

-
CANNIBAL
-
MISSIONARY

PROBLEM

Three

missionaries

and

three

cannibals

come

to

a

river

and

find

a

boat

that

holds

two
.

If

the

cannibals

ever

outnumber

the

missionaries

on

either

bank,

the

missionaries

will

be

eaten
.

How

shall

they

cross?

Here

comes

the

importance

of

knowledge
.

This

problem

can

although

be

solved

by

intelligent

algorithms

but

knowledge

plays

the

most

crucial

part


Need for formal languages



Consider

an

English

sentence

like
:

“The

boy

saw

a

girl

with

a

telescope”

Natural

languages

exhibit

ambiguity


Not

only

does

ambiguity

make

it

difficult

for

us

to

understand

what

is

the

intended

meaning

of

certain

phrases

and

sentences

but

also

makes

it

very

difficult

to

make

inferences


Symbolic

logic

is

a

syntactically

unambigious

knowledge

representation

language

(originally

developed

in

an

attempt

to

formalize

mathematical

reasoning)


KNOWLEDGE REPRESENTATION
TECHNIQUES IN AI


PROPOSITIONAL LOGIC

declarative statement

~
-
> Negation


-
> implication


-
> implies and implied by

v
-
> disjunction

^
-
> Conjunction

propositional logic

= sentences represent whole propositions

“2 is prime.” P

“I ate breakfast today.” Q


Syntax

syntax

= how a sentence looks like

Sentence
-
> AtomicSentence | ComplexSentence

AtomicSentence
-
> T(RUE) | F(ALSE) | Symbols

ComplexSentence
-
> ( Sentence ) | NOT Sentence |

Connective
-
> AND | OR | IMPLIES | EQUIV(ALENT)

Sentence Connective Sentence

Symbols
-
> P | Q | R | ...

Precedence: NOT AND OR IMPLIES EQUIVALENT

conjunction disjunction implication equivalence

negation


Semantics

semantics

= what a sentence means

interpretation:

assigns each symbol a truth value, either t(rue) or f(alse)

the truth value of T(RUE) is t(rue)

the truth value of F(ALSE) is f(alse)

truth tables (“compositional semantics”)

the meaning of a sentence is a function of the meaning of its
parts


Terminology



A sentence is
valid
if it is True under all possible assignments of

True/False to its propositional variables (e.g.
P
_:
P
)


Valid sentences are also referred to as
tautologies


A sentence is
satisfiable
if and only if there is some assignment
of

True/False to its propositional variables for which the sentence is

True


A sentence is
unsatisfiable
if and only if it is not satisfiable (e.g.

P
^:
P
)


Examples


either I go to the movies or I go swimming

2 is prime implies that 2 is even

2 is odd implies that 3 is even

(inclusive vs. exclusive OR)

(implication does not imply causality)

(false implies everything)


Semantic Networks

l Graph structures that encode taxonomic

knowledge of objects and their properties



objects represented as nodes



relations represented as labeled edges

l Inheritance = form of inference in which

subclasses inherit properties of superclasses


Frames


A

limitation

of

semantic

networks

is

that

additional

structure

is

often

necessary

to

distinguish



statements

about

an

object’s

relationships



properties

of

the

object

A

frame

is

a

node

with

additional

structure

that

facilitates

differentiating

relationships

between

objects

and

properties

of

objects
.

Called

a

“slot
-
and
-
filler”

representation


NORMAL Form in predicate LOGIC:


Rule
:
-

1
.


Replace

and

by

using

equivalent

formulas
.

2
.


Repeated

use

of

negation

~

(~

p)=F
.
Demorgan’s

law

to

bring

negation

in

front

of

each

atom
.

~

(GF)=

~G~F
.
Use

~x

F(x)=

x~F(x)

and

~xF(x)

=

x~F(x)


Then

use

all

the

equivalent

expressions

to

bring

the

quantities

in

front

of

the

expressions


Resolution in predicate LOGIC



i) R(a)

ii) R(x) M(x,b)

First replace a in place of x in 2
nd

premise and conclude
M(a,b).


e.g.


1.
Marcus was a man. Man (marcus)

2.
Marcus was a Pompeian. Pompeian (Marcus)

3.
Caesar was a ruler. Ruler (Caesar)



Nonmonotonic Reasoning



Collection

of

true

facts

never

decreases


Facts

changes

with

time


According

to

the

human

problem

solving

approach

the

truth

status

of

the

collected

facts

may

be

revised

based

on

contrary

evidences
.


Hence

the

nonmonotonic

reasoning

system

is

more

effective

in

many

practical

problems

solving

situations
.


Principles of NMRs



If

x

is

not

known,

then

conclude

y


If

x

cannot

be

proved,

then

conclude

y


e
.
g
.

1
:

To

build

a

program

that

generates

a

solution

to

a

fairly

a

simple

problem
.


e
.
g
.

2
:

To

find

out

a

time

at

which

three

busy

can

all

attain

a

meeting


dependency
-
directed

backtracking


Necessity of NMR



1.
The

presence

of

incomplete

information

requires

default

reasoning
.

2.
A

changing

world

must

be

decided

by

a

changing

database
.

3.
Generating

a

complete

solution

to

a

problem

may

require

temporary

assumption

about

partial

solution
.


Applications of AI


1.
PATTERN RECOGNISATION

2.
ROBOTICS

3.
NATURAL LANGUAGE PROCESSING

4.
ARTIFICIAL LIFE

5.
APPLICATIONS OF AI, BY INTELLIGENT ALGORITHMS


5.1 Mechanical translation


5.2 Game playing


5.3 Computer vision


5.4 Computer hearing


5.5 Creating original thoughts or works of art


5.6 Analogical thinking

Learning


Fundamental Problems of AI

1
.

The

ability

of

even

the

most

advanced

of

currently

existing

computer

systems

to

acquire

information

all

by

itself

is

still

extremely

limited
.

2
.


It

is

not

obvious

that

all

human

knowledge

is

encodable

in

“information

structures”

however

complex
.

e
.
g
.

A

human

may

know,

for

example,

just

what

kind

of

emotional

impact

touching

another

person’s

hand

will

have

both

on

the

other

person

and

on

himself
.


3
.


The

hand
-
touching

example

will

do

here

too,

there

are

some

things

people

come

to

know

only

as

a

consequence

of

having

been

treated

as

human

beings

by

other

human

beings
.

4
.


The

kinds

of

knowledge

that

appear

superficially

to

be

communicable

from

one

human

being

to

another

in

language

alone

are

in

fact

not

altogether

so

communicable


Thank You!!!