John Pollock directs the OSCAR Project, funded in part
by the National Science Foundation. The goal of the OSCAR Project is the formulation
of a general theory of rationality and its implementation in an artificial
rational agent. The project is predicated on the view that philosophy has
an essential role to play in artificial intelligence. The function of artificial
agents is to draw conclusions and make decisions on the basis of information
supplied to them. But we do not want them to draw just any old conclusions
or make just any old decisions. We want them to draw rational conclusions
and make rational decisions. We can make a limited amount of progress
in the task of building such an agent by relying upon our untutored intuitions
about what is rational. But to build a general purpose rational agent, we
need a general account of how the agent is to behave -- a theory of rationality.
The construction and implementation of such a theory is the objective of the
OSCAR Project.
OSCAR is a fully implemented architecture for rational agents,
based upon a general purpose defeasible reasoner. The defeasible reasoner
is the world's first such reasoner capable of operating in a rich logical
environment like first-order logic in which logical consistency is not decidable.
The theory underlying OSCAR is described in Cognitive Carpentry (Bradford/MIT
Press, 1995). Click here for an overview
of OSCAR, along with links allowing the current version of OSCAR to be
downloaded, along with The OSCAR Manual. The latter details the construction
of OSCAR, and explains its application to particular reasoning problems.
