Artificial Intelligence

what is knowledge ?
data that
knowledge refers to the information or

taks.
machine was to make decisions or perform

For example if you want ai to recognize picture

,
of
show it lot of pictures of cate so
cats you would
,

like Here the

that it can what a cat looks
learn
.

for Al
picture of cats are
the knowledge our
 Representation Issues
knowledge
which attributes are
1
. Important Attributes: Deciding
crucial . It's like
most important to represent is
the most important features of an
object
nicking
to describe it accurately
how
Relationship among attributes Representing
:
·
2
.
different attributes are related to each other
It's like understanding how different parts of an
to make it a whole
object fit together
on the level of detail
.
3 Choosing Granularity .
Deciding
to represent is crucial . It like
deciding whether
to look at the picture up clone or from far away.
which objects to include
s Set of objects
:
Deciding
It's like
in
knowledge representation is important
.

deciding what examples towe to teach a concept .

S .

Finding the right structure :

Deciding the best

way
to structure the knowledge representation
crucial. It's
is like deciding how to
organize
information in a book to make it eary to
understand.
REPRESENTATION I MAPPING

Representation refers to how facts are represented in

Al
system. This is how you would represent a
fact in your mind or paper
For example the fact
.
,

"the cat is on the mat" could be represented

in different ways such as

Symbolic Representation using symbols and

Logic
represent the such "cat-on (mat)
to fact as

natural
Semantic Representation : wing language
represent the such "The cat is on the
to
fact as

matt ?

Statistical Representation : wing probabilities to

represent the fact ,
such as "The probablity of
the cat
being on mat is 0 8 %.
.
APPROACHES TO KNOWLEDGE REPRESENTATION

Relational Approach :

In this approach, knowledge is represented in terms

of relationships between objects For example we

.
,

could represent the fact that "Tom is the owner of

a black callwing a
graph with nodes for Tom
Lat
and "Black ,
and
edge between them represent
the "owner" relationship
.
ing
-

Example :
Owner Pet color
Tom cat Black

Brown
mary Dog
Glunk Cat orange
Inheritable Representation :
In this approach objects are represented as
hierarchies ,
where each objects inherits attributes
and relationships from it's parent objects.

Example : we could represent "Black Cats as

the child object of Cath which is child

,

object of "Animal" Black Cate

a
.
That means
would in herit attributes from both "Carr and
6 Animal

Animal

↑
Cats
Tr

Black catGrange cat

Inferential Representation :
In this approach knowledge
,
is represented as
statements be wed to
rules or
logical that can

infer new
knowledge
Example : if we know thatAll cate are mammals
and "Black cat is a cat we can infer that
,

"Black cat is a mammal This

inference can

be represented as a rule or
logical statement.
such as "if an animal is a cat and all cuts are
mammals ,
then the animal is mammal ?
DECLARATIVE/ PROCEDURAL REPRESENTATION :

In this approach declarative

,
knowledge is

represented separately from procedural

knowledge
.
Example : We could represent the declarative

knowledge of "What cats eaty as a set of

se
facts ,Suchascate Carnivor
and

are the

procedural knowledge of "how to

feed a cate
separately ,
such as "put the
food in a bow
"place the bowl the :
and on
floor
 PREDICATE LOGIC

Predicate
Logic is a
type of formal logic that
deals with relationship between objects ,
and
it allows us express complex ideas in
to

a precise and rigorous way . It's often und

in Al to represent knowledge and make

inferences

variables :
In first-order Logic ,
variables are

symbols that stand for objects in a domain of

discourse. They are usually denoted by lowercare
letters such as x, 2 Etz.
y , .

Predicates : Predicates are statements that define

relationship between the
objects They
.
are
usually
denoted in uppercare letters such as P C R ets. , , ,

example : "is a cate could be represented by

Cat(X) is a variable that
predicate ,
where X

.
stands for object in a demain
Functions Functions
: are
symbols thatrepresent
operations that take one or more object as output.
They are
usually denoted by lowercare letters
such as f gih
, .
Etc .

I
Example : The
function "add could be represented
by symbol +.
Connectives :
that allow
Connectives are
symbols
us to combine predicates and make more

complex statements
. The most
common connectives
in
first logic "and"(1)
order "
or (V). are
,

"not"171 and "if then "ks) " if and only if ,

((3)

Equality
"
: "is used to indicate
symbol
The ==

that two objects are equal .

For example ,
6x ==
yo
means that the variable a and variable y represent
same object in the demain.

Quantifiers allows us to express how

Quantifiers :
many objects satisfy
a
predicate The most common .

quantifiers are ferall"(V) and "existe(7) "Forall .

is read as for all" and means that a predicate

for every object in the remain
.
is true
%
All cats are mammals" :

-x (at(x) > Mammal (x)

-
Read flw and blu reasoning arine