Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 4

NAME: SAMUEL GACHARI

REG NO: HDB212-0564/2017

COURSE: BBIT 4.2

UNIT: ARTIFICIAL INTELLIGENCE

ASSIGNMENT
QUESTION 2
Recall the semantic network about baseball players from the lectures and Rich &
Knight:How could the same information could be represented in natural language? Or in
terms of first order logic? Or as a database? Is one form of representation clearly better
than the others, or are there advantages and disadvantages to each? Are there aspects of
baseball players that none of these forms of representation can accommodate.

Natural Language
 Three-Finger Brown’s team is Chicago Cubs
 Three-Finger Brown is a pitcher
 Three-Finger Brown’s batting-average is 106
 Three-Finger Brown is a Baseball player
 Three-Finger Brown is an adult male
 Three-Finger Brown is right handed

First order logic


 Pitcher(Three-Finger Brown)
 Baseball-player(Three-Finger Brown, Pee-Wee-Reese)

Database
Baseball-Player = { name : max 32 characters
height : 3 digits in range 100-300
type : Pitcher or Fielder
team : Chicago-Cubs, Brooklyn Dodgers
}

There are clearly many more representational formalisms that might be useful. For a start, we
have only really considered symbolic representations. There also exist non-symbolic (e.g.
pictorial) representations. So-called sub-symbolic representations are also possible (e.g. as one
finds in the activation patterns of neural network systems). In selecting a representational
formalism one needs to consider exactly what has to be represented, and how that knowledge
needs to be processed. We should also consider how the system will go about acquiring new
knowledge. In practice, there is no single best knowledge representation system that can be used
for everything. In building large complex AI systems, one will usually want to employ many
different types of knowledge representation.

QUESTION 3
List two advantages and two disadvantages of using each of the following as a general
purpose AI knowledge representation scheme:
(a) Natural Language
Advantages
1. relieves burden of learning syntax
2. not training
disadvantages
1. requires clarification dialogue
2. may not show context
(b) Standard Computer Science Databases
Advantages
1. duplication of same data
2. wastage of storage space
disadvantages
1. database complexity
2. damage to database affects virtually all application programs
(c) Frames
Advantages
1. allows the user to view multiple documents within a single web page
2. it is possible to load pages from different servers in a single frameset.
Disadvantages
1. frames can make the production of a website complicated
2. users have become to familiar with normal navigation using tables and so on, that
navigating through a site that uses frames can be a problem.
(d) Semantic Networks
Advantages
1. the semantic network is more natural than the logical representation.
2. They are simple and can be easily implemented and understood.
Disadvantages
1. there is no standard definition for link names
2. links on object represent only binary relations
(e) First Order Logic
Advantages
1. it is very expressive
2. it has unambiguous syntax and semantic
(f) If-Then Rule Based Systems
Advantages
1. these systems are very expressive
2. the rules lead to a degree of modularity
disadvantages
1. there is a lack of precise semantics for the rules
2. the systems are not always efficient

QUESTION 5
What other AI knowledge representation schemes exist, apart from those listed in the
above
questions 3 and 4?
1. Relational Knowledge
2. Inheritable Knowledge
3. Inferential Knowledge
4. Declarative Knowledge
5. Procedural Knowledge

You might also like