Procedural Knowledge Representation in Soar
Soar's knowledge representation is not transparent but rather like a
black box. This happens because the representation is not
declarative but process-like.
Long-term memory (productions) can not
be examined directly by the architecture: the only way to determine the
contents of a production is for it to fire. Thus, long-term memory
is impenetrable, meaning, in this case, that Soar can not reason
about how to change the content of individual productions. Since
productions (and
chunks) are permanent and
may not be examined or changed, Soar performs best in domains
that are
consistent and experiences a great deal of
difficulty domains which
vary continuously.
To go to the Soar property list, press UP
To return, press HOME.
To go to the next document, press NEXT.