Page 159 - DCAP310_INTRODUCTION_TO_ARTIFICIAL_INTELLIGENCE_AND_EXPERT_SYSTEMS

Unit 8: Structured Representation of Knowledge

compare them. Data organization in predicate logic consists of a set of unambiguous constants Notes
(representing entities in the domain), a set of unambiguous predicates (representing relations
between entities in the domain), a set of functions (mapping between sets), variables, quantifiers,
and logical connectives. Inference in predicate logic is achieved by applying deductive inference
rules, e.g. by means of resolution. For an introduction to the logical approach to knowledge
representation.
A practical computer language based on a limited version of predicate logic is PROLOG. Below
is a small program that expresses the fact that Socrates and Plato are human, and the rule that if
x is human, then x is mortal:
human(socrates).
human(plato).
mortal(X):- human(X).
The interpreter of PROLOG uses these facts and rules to derive other facts. For example, the
following dialog is possible, where we ask whether Socrates and Descartes are mortal, and who
are all mortal beings the system knows. The system infers, for instance, that Socrates and Plato
are mortal. Notice that PROLOG gives a negative answer for everything that does not occur in
the knowledge base.
|?-mortal(socrates).
yes
|?-mortal(descartes).
no
|?-mortal(X).
X=socrates;
X=plato;
No

Predicate logic has some severe limitations as a tool for representing linguistic knowledge
which is incomplete, inconsistent, dynamically changing, or relating to time, action and beliefs.

!
Caution For all these problems, special purpose logics are being designed. An example is
default logic, which handles exceptional information without having to modify existing
general knowledge.

Self Assessment

State whether the following statements are true or false:
5. The interpreter of PROLOG uses these facts and rules to derive other facts.
6. Logic has been used as a formal foundation for knowledge representation in AI.

8.4 Grammar Formalisms

Grammar formalisms constitute a special type of formalism for natural language processing,
even though they are not unrelated to the knowledge representation paradigms and formalisms
discussed earlier. They often use a different terminology, due to the different background of the
developers, which is linguistics, logic, and theoretical computer science rather than AI, and use
special notations for linguistic strings and structures. Most grammar formalisms were developed
as part of the efforts to build systems for natural language understanding, which up to now
received more attention in AI than natural language generation.

LOVELY PROFESSIONAL UNIVERSITY 153