KNOWLEDGE

REPRESENTATION
KNOWLEDGE REPRESENTATION AND
MAPPING
 To solve the complex problems encountered in artificial
intelligence ,one needs both a large amount of
knowledge and some mechanisms for manipulating that
knowledge to create solution to new problem.
 Ways of representing knowledge have been exploited in
AI programs, dealing with two different kind of entities:
1. Facts: truths in some relevant world. These are the
things we want to represent.
2. Representations of facts in some chosen formalism.
These are the things we will actually be able to
manipulate.
KNOWLEDGE AND MAPPING CONT…..
 Structuring these entities is as two levels:
1. The knowledge level , at which facts(including each agent’s
behaviors and current goals) are described.
2. The symbol level, at which representation of objects at
knowledge level are defined in terms of symbols that can be
manipulated by programs
MAPPING B/W FACTS &
REPRESENTATIONS
*

Reasoning Program

Facts
Internal* Representations

English English
Understanding Generation

English Representation

representation mappings
MAPPING B/W FACTS &
REPRESENTATIONS
 In the above figure all the links are called representation
mappings.
 The forward representation mapping maps from facts to
representations.
 The backward representation mapping goes the other
way from representations to facts.
 desired real reasoning
Initial facts Final facts

* forward backward *
representation representation
mapping mapping

Internal representation of Internal representation of

initial facts final facts
operation
of program

Figure: Representation of Facts

KNOWLEDGE REPRESENTATION
SYSTEM
 A good system for the representation of knowledge in a
particular domain should posses the following four
properties:
1) Representational Adequacy –the ability to represent all of
the kinds of knowledge that are needed in that domain.
2) Inferential Adequacy- the ability to manipulate the
representational structures in such a way as to derive new
structure corresponding to new knowledge inferred from
old.
3) Inferential Efficiency- the ability to incorporate into the
knowledge structure additional information that can be
used to focus the attention of the inference mechanism in
the most promising direction.
KNOWLEDGE REPRESENTATION SYSTEM
CONTD…
4) Acquisitional Efficiency- the ability to acquire new
information easily. The simplest case involves direct
insertion , by a person , of new knowledge into database.
Ideally , the program itself would be able to control
knowledge acquisition.
• Unfortunately , no single system that optimizes all of the
capabilities for all kind of knowledge has yet been found.
• As a result multiple knowledge representation system exists.
• Most of the programs rely on more than one knowledge
representation technique.
 APPROACHES TO KNOWLEDGE
REPRESENTATION
 Following four approaches to represent knowledge is followed:
1) Simple Relational Knowledge
2) Inheritable Knowledge
3) Inferential Knowledge
4) Procedural Knowledge
 SIMPLE RELATIONAL KNOWLEDGE
 Represent declarative facts as a set of relations of the same
sort used in database systems.
 It is simple because standing alone, it provides very weak
inferential capabilities.
 But knowledge represented may serve as input to inference
engines.
 Providing support for relational knowledge is what
database systems are designed to do.
 The practical issues that arise in linking a database system
to a knowledge representation system have already been
solved in commercial products.
 So these structures are not further explored in Artificial
Intelligence.
 SIMPLE RELATIONAL KNOWLEDGE

Player Height Weight Bats-Throws

Hank Aaron 6-0 180 Right-Right
Willie Mays 5-10 170 Right-Right
Babe Ruth 6-2 215 Left-Left
Ted Williams 6-3 205 Left-Right
Player-info(Hank Aaron, 6-0,180,right-right)
 INHERITABLE KNOWLEDGE
 Knowledge about objects, their attributes, and their values
represented in simple relational knowledge need not to be as simple.
 The basic representation must be augmented with inference
mechanism that operate on the structure of the representation.
 One of the most useful form of inference is property inheritance, in
which elements of specific classes inherit attributes and values from
more general classes in which the knowledge is represented.
 The knowledge on which property inheritance can be applied is
inheritable knowledge and the knowledge structures for inheritable
knowledge (which supports property inheritance) are called slot-and-
filler structures.
 Weak Slot-and-filler Structures includes semantic network or a
collection of frames.
 Strong Slot-and-filler Structures includes Conceptual Dependency
and Scripts.
INHERITABLE KNOWLEDGE CONTD..
 In order to support property inheritance by the structures, the
objects must be organized into classes and classes must be
arranged in generalization hierarchy.
 For example, in semantic networks, boxed nodes represent
objects, classes and values of attributes of objects, lines
represents attributes.
 The arrows on the lines point from an object to its value along
the corresponding attribute line.
 Two important attributes in these structures are:
1) instance-It means class membership. A instance B means
object A is an instance of Class B.
2) isa- It means class inclusion. A is a B means class A is
included in class B.
 Person Right
handed
isa

Adult-male 5-10
height

isa 6-1
height

bats Baseball
Equal to
player Batting-average
handed
.252
isa isa

Batting-average Batting-average
.106 Pitcher Fielder .262

instance instance

Chicago team Three-finger- Pee-Wee- team Brooklyn-

cubs Brown Reese dodgers
 INHERITABLE KNOWLEDGE CONTD..
 The knowledge represented in the previous slide can also be
represented as a collection of frames.
 Each individual frame represents the collection of attributes
and values associated with a particular node:
Baseball-player
isa: adult-male
bats: (EQUAL handed)
height: 6-1
batting-average: .252
Viewing a Node as a Frame
 INFERENTIAL KNOWLEDGE
 Generates new information .
 Generates new information from the given information.

This new information does not require further data gathering form
source, but does require analysis of the given information to
generate new knowledge.
− Given a set of relations and values, one may infer other values or
relations.
− In addition to algebraic relations, a predicate logic (mathematical
deduction) is used to infer from a set of attributes.
− Inference through predicate logic uses a set of logical operations to
relate individual data.
The symbols used for the logic operations are : " → "
(implication), " ¬ " (not), " V " (or), " Λ " (and), " ∀ " (for all), " ∃
" (there exists).
INFERENTIAL KNOWLEDGE
CONTD….
 Examples of predicate logic statements :
1. Wonder is a name of a dog : dog (wonder)
2. All dogs belong to the class of animals : ∀ x : dog (x) →
animal(x)
3. All animals either live on land or in water : ∀ x : animal(x) →
live (x, land) V live (x, water)

We can infer from these three statements that :

" Wonder lives either on land or on water.“

Procedures like forward chaining, backward chaining and

resolution ( proof by contradiction / refutation) are used to
infer new knowledge from the inferential knowledge base.
 PROCEDURAL KNOWLEDGE
 Here the knowledge is a mapping process between domains that
specifies “what to do when” and the representation is of “how to
make it” rather than “what it is”.
 This is represented as small programs that know how to do specific
things, how to proceed.
 Example : a parser in a natural language has the knowledge that a
noun phrase may contain articles, adjectives and nouns. It thus
accordingly call routines that know how to process articles,
adjectives and nouns.
 The procedural knowledge : − may have inferential efficiency, but
no inferential adequacy (it is very difficult to write a program that
can reason about other program’s behavior) and acquisitional
efficiency ( the process of updating large piece of code is difficult).