Page 89 - DCAP506_ARTIFICIAL_INTELLIGENCE

Unit 6: Knowledge Representation using Rules

The apple appears red. Then the consequence: Notes
The apple is red follows as a goal, since there is no clause whose conclusion matches the sub-goal
the apple is enlightened by a red light, and thus the two conditions for the only clause that can
be utilized in solving the goal both hold. Though, provided the additional clause the apple is
illuminated by a red light, the sub-goal now accomplish something and the top-level goal now
fails, non-monotonically withdrawing the consequence the apple is red. Though, the consequence
not the apple is red now succeeds instead.

Goals and conditions of clauses can be generalized from conjunctions of literals to arbitrary
formulae of first-order logic. The simplest manner to perform so is to use auxiliary predicates
and clauses (Lloyd and Topor, 1984).

Example: The goal:

Demonstrate that for all exams, if the exam is a final exam, then you can revise for the exam in
the library can be converted into the normal logic programming form:
Demonstrate that not the library is useless; the library is futile if the exam is a final exam and not
you can study for the exam in the library.

!

Caution The conversion applies in logic programming only when the conditional is inferred
as a goal, and not when it is inferred as a clause.
The computational benefit of the conversion is that it decreases the problem of identifying
whether a random sentence of first-order logic holds with respect to a specified logic program
to the two simple inference rules of backward reasoning and negation as breakdown alone.

Notes Horn clauses are named after the logician Alfred Horn, who studied some of their
model-theoretic properties.

Comparison between Backward and Forward reasoning is discussed below.
Self Assessment

Fill in the blanks:
7. ............................ contains three major classes of application: as a general-purpose
programming language, a database language, and a knowledge representation language
in AI.
8. As a database language, logic programming simplifies ............................ databases, to
comprise general clauses as well as facts.
9. Conditionals in logic programs are also known as ............................ .
10. ............................ are the particular case where all of the conditions are atomic formulae.

6.3 Forward vs Backward Reasoning

A search procedure must locate a path among initial and goal states. There are two directions in
which a search process could progress.

LOVELY PROFESSIONAL UNIVERSITY 83