Structured Knowledge Representation in AI

Representation(CsIt Nepal)

A representation consists of two sets and a mapping between them. The elements of each set are objects, relations, classes, laws, actions. The first set is called the represented domain, and the second one is called the representation domain.

This mapping allows one to reason about the represented domain by performing reasoning processes in the representation domain and transferring the conclusions back into the represented domain by using reverse mapping between the two sets.

The processes which manipulate the objects and the relationships in the representation domain are also part of the representation.

Knowledge representation, which is concerned with encoding the human knowledge into the form that is efficiently manipulated by the computer. The different kinds of knowledge that is represented in AI systems are ;-

• Objects: Guitar has Strings.
• Events: Ram played Guitar.
• Performance: Dirt cleaned by vacuum cleaner
• Meta-Knowledge: Ram knows that she can read street signs along the way to find where it is
• Facts: Truths about real world and what we represent

Knowledge Representation (CsIt Nepal)

In order to solve a problem, it is necessary to know a large amount of knowledge and some mechanisms for manipulating that knowledge to create solutions to new problems. The following two entities focus to all discussion of representation:

• Facts: are truths in some applicable world. These are the things required to represent. In other words, fact level refers to the set of possible states of affairs.
• Representation of facts in some formal procedure. These are the things which are to be manipulated.

One way to visualize these entities is as two levels:

• The knowledge level, at which facts are described.
• The symbol level, at which representation of objects at the knowledge level are defined in terms of symbols, which can be manipulated by programs.

Approaches to Knowledge Representation(CsIt Nepal)

For a good system, the representation of knowledge in a particular field should possess the following four properties:

• Representational Adequacy: - the ability to represent all of the kinds of knowledge that are needed in that field.
• Inferential Adequacy: - the ability to manipulate the representational structures in such a way as to derive new structures corresponding to new knowledge.
• Inferential Efficiency: - the ability to incorporate additional information into the knowledge structure that can be used to focus the attention of the inference mechanisms in the most promising directions.
• Acquisitional Efficiency: - the ability to acquire new information easily. This involves a direct insertion, by a person, of a new knowledge into the database.

Formal logic-connectives:

In logic, a logical connective which is also called a logical operator is a symbol or word that is used to connect two or more sentences (of either a formal or a natural language) in a grammatically valid way, so that the compound sentence produced has a truth value dependent on the respective truth values of the original sentences.

Each one of the logical connective can be expressed as a function, called a truth function. For this reason, logical connectives are sometimes called truth-functional connectives.

Commonly used logical connectives include:

• Negation (not) (¬ or ~)
• Conjunction (and) (^ , &, or )
• Disjunction (V or v)
• Material implication (if...then)
• Biconditional (if and only if) (iff) (xnor)

For example, the meaning of the two statements it is raining and I am indoors is transformed when the two are combined together with logical connectives:

• It is raining and I am indoors (P ^ Q)
• If it is raining, then I am indoors (P à Q)
• It is raining if I am indoors (Q à P)
• It is raining if and only if I am indoors (P ↔ Q)
• It is not raining (¬P)

Given statement P = It is raining and Q = I am indoors.

Issues in Knowledge Representation

• Are any attributes of objects so basic that they occur in almost every problem domain? If there are, it is necessary to make sure that whether they are handled appropriately in each of the proposed mechanism. If such attributes exist, what are they?
• Are there any important relationships that exist among attributes of objects?
• At what level should knowledge be represented? Is there a good set of primitives into which all knowledge can be broken done? Is it helpful to use such primitives?
• How should sets of objects be represented?
• Given a large amount of knowledge stored in a database, how can relevant parts be accessed when they are needed?

Semantic Network (CsIt Nepal)

The fundamental of semantic nets is that the meaning of a concept comes from the ways in which it is connected to other concepts. Information is represented as a set of nodes connected to each other by a set of labeled arcs in a semantic net that represent the relationships among the nodes.

To find relationships among objects by spreading out from each of two nodes and seeing where the activation met, the semantic net is used which process is called intersection search. Through this process, it is possible to use the network of the above figure to answer such questions as "What is the connection between Nepal and Red"? To answer more structured questions, it requires networks that are themselves more highly structured.

Representing Nonbinary Predicates

Semantic nets are a natural way to represent relationships that would appear as ground instances of binary predicates in predicate logic. For example, some of the arcs from the above figure could be represented in logic as: is a(Person, Mammal) instance (MMC, Person) team (MMC, Nepal) uniform color(MMC, Red).

Frames (CsIt Nepal)

A collection of attributes i.e. generally called slots and associated values that describe some entity in the world is known as a frame. A frame describes an entity in some absolute sense. It also represents the entity from a particular point of view. A single frame taken along is not often useful. Instead, a frame system is build form the collection of frames are connected to each other by virtue of the fact that the value of an attribute of one frame may be another frame.

Frames as Sets and Instances

Person

Is a: Mammal *

Handed: Right

Adult-male

Is a: Person *

Height: 5-6

Cricketer

Is a: Adult-male *

Height: 5-6 *

Bats: right-handed

Batsman

Is a: Cricketer *

Batting average: 48

Nepal Under-19

Instance: Cricketer

Height: 5-6

Bats: right batting

Average: 48

Team: Nepal

Uniform-color: Red

Nepal

Instance: Nepal Under-19

Team-size: 11

Set theory gives a good basis for understanding frame systems. All frames represents either a class or an instance. The above frame system is slightly modified than that of the previous network. Here all the classes are the frames Person, Adult-Male, Cricketer and Batsman. The frames Nepal Under-19 and Nepal instances. The is a relation that has been used without a precise definition is, in fact, the subset relation. The set of adult-males is a set of the people. The set of cricketers is a subset of adult males, and so forth. So, instance relation corresponds to the relation element-of. Nepal Under19 is an element of the set of Cricketer. Both the instance relations have inverse attributes, which are known as subclasses and all-instances. A class represents a set; there are two types of attributes that can be assigned with it. The attributes about the set itself and the attributes that are to be inherited by each element of the set. In the preceding example, they are distinguished by prefixing the former with "*".

Conceptual Dependencies:(CsIt Nepal)

Conceptual Dependency is originally developed to represent knowledge that is acquired from natural language input.

The goals of Conceptual Dependency are:

• To provide help in the drawing of inference from sentences.
• To be independent of the words used in the original input.
• That is to say: For any 2 or more sentences that are identical in meaning there should be only one representation of that meaning.

Conceptual Dependency provides:

• IT provides a structure into which nodes representing information can be placed.
• It provides a specific set of primitives

Using both abstract and real physical situations sentences are represented as a series of diagrams depicting.

Conceptual Dependency (CD) is a theory which represents the kind of knowledge about the events which is typically contained in natural language processing sentences. The goal is that to represent the knowledge in a way that facilitates drawing the inferences from the natural language sentences.

CD representation of a sentence is built not out of primitives corresponding to the words in the sentence because of the two concerns mentioned but rather out of conceptual primitives which can be combined to form the semantic meanings of the words in any language. At first, the conceptual dependency theory was developed by Schank in 1973 and was further developed in 1975 by the same author. It has been implemented in different programs which read and understand natural language text processing. At a particular level of granularity unlike semantic nets that gives only a structure into which nodes representing information and a specific set of primitives, out of which representations of particular pieces of information can be developed.

In CD’s the symbols have the following meanings.

Arrows indicate the decision of dependency.

Double arrow indicates two-way link between actor and action

P indicates past tense.

ATRANS is one of the primitive symbol acts used by the CD theory which indicates transfer of possession

O indicates the object case relation

R indicates the recipient case relation.

Scripts(CsIt Nepal)

A structure which prescribes a set of circumstances which could be expected to follow on from one another is a script. Scripts are used for representing knowledge about common sequences of events.

It is similar to a thought sequence or a chain of situations that could be anticipated.

It could be considered to consist of a number of slots or frames but with more specialized roles.

Function of Scripts

• Events to occur in known runs or patterns.
• Relationships between events exist.
• Entry conditions exist that allow an event to take place
• Prerequisites exist for events taking place.

The components of a script include:

Entry Conditions

These must be satisfied before an event in the script occurs.

• Results: Results are the conditions which will be true after events in script occur.
• Props: Slots represent objects that are involved in events.
• Roles: Persons who are involved in the events.
• Track: Variations on the script. Various tracks may share components of the same script.
• Scenes: The sequence of events which occur. Events are represented in conceptual dependency

The classic example is the restaurant script:

Scene: A restaurant with an entrance and tables.

Actors: The diners, servers, chef.

Props: The table setting, menu, table, chair.

Acts: Entry, Seating, Ordering a meal, Serving a meal, Eating the meal, requesting the check, paying, leaving.

Advantages of Scripts:

• It has the ability to predict events.
• A single coherent interpretation builds up from a collection of observations.

Disadvantages:

• It is less general than frames.
• It may not be suitable to represent all kinds of knowledge.

REFERENCE

CsIt Nepal. (n.d.). Retrieved from csitnepal.com: http://csitnepal.com/elibrary/notes/

Knight., E. R. (n.d.). Artificial Intelligence. In E. R. Knight.. The McGraw-Hill Companies.

Longman, J. F. (n.d.). Computer Netwroking: A top-down approach featuring the Internet. In a. F. Longman, Computer Netwroking: A top-down appriach featuring the Internet (p. 65).

Patterson., D. W. (n.d.). Introduction to Artificial Intelligence and Expert Systems. In D. W. Patterson., Introduction to Artificial Intelligence and Expert Systems. Prentice Hall.

slideshare. (n.d.). Retrieved from www.slideshare.net: http://www.slideshare.net/u053675/artificial-intelligence-1419854

Stuart Russell, P. N. (n.d.). Artificial intelligence : a modern approach. In P. N. Stuart Russell, Artificial intelligence : a modern approach. pearson.

tutorialspoint. (n.d.). Retrieved from www.tutorialspoint.com: http://www.tutorialspoint.com/artificial_intelligence/artificial_intelligence_tutorial.pdf

Winston, P. H. (n.d.). Artificial Intelligence. In P. H. Winston, Artificial Intelligence. addison wesley.