Page 170 - DCAP506_ARTIFICIAL_INTELLIGENCE

Artificial Intelligence

Notes Self Assessment

Fill in the blanks:
8. An ......................... system accepts an example (i.e. a training example) and illustrates what
it learns from the example.

9. ......................... is a system that incorporates problem solving, planning, and learning
methods in a single design.

12.6 Learning by Parameter Adjustment

Here the learning system depends on evaluation procedure that merges information from
numerous sources into a single summary static.

Example: The factors like demand and production capacity may be merged into a single
score to signify the likelihood for increase of production. But it is hard to know a priori how
much weight should be associated to each factor.
The accurate weight can be located by taking some approximation of the correct settings and
then permit the program alter its settings based on its experience. This type of learning systems
is functional when little knowledge is obtainable.

Example: In game programs, the factors like piece benefit and mobility are merged into
a score to decide whether a specific board position is desirable. This single score is nothing but
a knowledge which the program collected through calculation.

Self Assessment

Fill in the blank:
10. In case of learning by........................., learning system depends on evaluation procedure
that merges information from numerous sources into a single summary static.

12.7 Learning with Macro-operators

Series of actions that can be considered as a whole are known as macro-operators. Once a
problem is solved, the learning component takes the calculated plan and accumulates it as a
macro-operator. The requirements are the initial conditions of the problem just solved, and its
post conditions matches to the goal just attained.
The problem solver competently accesses the knowledge base it gained from its preceding
experiences. By generalizing macro-operators the problem solver can even crack diverse
problems.

!

Caution Generalization is performed by substituting all the constants in the macro-operators
with variables.

Example: The STRIPS, is a planning algorithm that engaged macro-operators in it’s
learning segment. It constructs a macro operator MACROP, that comprises preconditions,

164 LOVELY PROFESSIONAL UNIVERSITY