Meta-Reasoning Architecture: Domain Knowledge
Meta-Reasoning Architecture: Domain Knowledge
MAX may solve problems in many domains, any one of which can be defined
along with the appropriate primitive operators in the base domain area
of the architecture definition, encoded in the architecture's
uniform
lframe representation.
General
behaviors are thought to be
useful for general intelligence, while
base domain knowledge and operations specifically assist problem solving for
particular agents, including domain-specific
learning and heuristics.
As indicated by the block in the
graphical picture of MAX,
additional knowledge may be added to MAX, denoting
the extensibility of the architecture. The
references
examined for this document
did not utilize this aspect of the architecture nor indicate how it might
be used. Note that this additional knowledge
cannot overcome the limitations of the
declarative
representation of knowledge,
including the monotonic dependence of efficiency on the retention of
new knowledge, nor the limitations of two-valued
logic on the ability to represent continuous variables, especially
continuous time.
Return to the top of this architecture.
Current Location: Meta-Reasoning Architecture - Domain Knowledge
