Page 171 - DCAP506_ARTIFICIAL_INTELLIGENCE

Unit 12: Learning

post-conditions and the series of actions. The macro operator will be accessed in the upcoming Notes
operation.

Self Assessment

Fill in the blanks:

11. Series of actions that can be considered as a whole are known as ......................... .
12. Generalization is performed by substituting all the ......................... in the macro-operators
with variables.

12.8 Learning by Chunking

Chunking is comparable to learning by means of macro-operators. Usually, it is accessed by
problem solver systems that utilize production systems.
A production system comprise of a set of rules that are in if-then outline. That is specified a
specific situation, what are the actions to be achieved.

Example: If it is raining then take umbrella.
Production system also comprises knowledge base, control strategy and a rule applier. To solve
a problem, a system will contrast the present condition with the left hand side of the rules.
If there is a match then the system will achieve the actions illustrated in the right hand side of the
corresponding rule.

Problem solvers explain problems by applying the rules. Some of these rules may be more
functional than others and the results are accumulated as a chunk. Chunking can be utilized to
study common search control knowledge. Numerous chunks may encode a single macro-operator
and one chunk may contribute in a number of macro sequences. Chunks learned in the starting
of problem solving, may be used in the afterward stage. The system maintains the chunk to
utilize it in solving other problems.
Soar is a common cognitive architecture for producing intelligent systems. Soar needs knowledge
to solve several problems. It obtains knowledge by means of chunking mechanism. The system
learns reflexively when impasses have been resolved. An impasse occurs when the system does
not have enough knowledge. Thus, Soar selects a new problem space (set of states and the
operators that influence the states) in an offer to resolve the standoff. While resolving the
impasse, the individual steps of the task plan are collected into bigger steps called chunks. The
chunks reduce the problem space search and so augment the efficiency of achieving the task.

In Soar, the knowledge is accumulated in long-term memory. Soar utilizes the chunking
mechanism to generate productions that are accumulated in long-term memory. A chunk is
nothing but a great production that does the work of the whole sequence of smaller ones. The
productions have a set of conditions or patterns to be matched to functioning memory which
comprise of existing goals, problem spaces, states and operators and a set of actions to carry out
when the production fires. Chunks are generalized before accumulating. When the same impasse
appears again, the chunks so gathered can be utilized to resolve it.
Self Assessment

Fill in the blanks:
13. A ......................... system comprise of a set of rules that are in if-then outline.

LOVELY PROFESSIONAL UNIVERSITY 165