Adding New Knowledge in Soar
Adding New Knowledge in Soar
Because all of Soar's long-term memory
is encoded as productions, new knowledge may be added to Soar simply
by coding new productions. This ability is facilitated by Soar's uniform and uncompartmentalized
knowledge representation and the absence of an upper limit on the
size of the knowledge base.
Although these productions are structurally unmodular, problem spaces provide an inherent
structural organization for the addition of productions. This is
important for building large Soar systems, such as those which act as
expert systems.
The addition of new knowledge in Soar brings up two criticisms related to
the architecture's representation of knowledge:
- Results Highly Dependent Upon Choice of Representation
This is an issue for all cognitive architectures that use some form of
symbolic representation.
For Soar, both the efficiency and generality of the results of problem solving are critically
dependent upon the choice of representation when the system is first
designed.
- Fixed and Immutable Representation
The choice of
representation is not easily changed. In particular, the impenetrability of long-term memory and the permanence of
these productions make adapting the representation very difficult.
Return to the Issues Index for this Architecture.
Return to the top of this architecture.
Go to a Discussion of this
issue for multiple architectures.
Current Location: Soar-Issues-Adding New Knowledge