Self Reflection in Prodigy
Self Reflection in Prodigy
Prodigy incorporates several different learning modules which make use of the
agent's
declarative,
glass-box representations to
reflect on the agent's actions and knowledge:
- STATIC accesses the contents of
the agent's knowledge base to form new control rules
- EBL and Analogy
use problem-solving traces to create new control rules based on the
agent's previous actions
The action of these modules requires both a
static environment and which is
consistent throughout both
problem solving and this self reflection. Without a consistent
and stable base, the modules' inspection of the problem solving
behavior could not guarantee appropriate
generalization or even
correct learning during reflection.
Return to the top of this architecture.
Go to a discussion of this capability
for multiple architectures.
Current Location: Prodigy-Capabilities-Self Reflection