Folk Psychology for Human Modelling: Extending the BDI Paradigm
Emma Norling
Friday, 28th May 2004
0930 - 1100
Abstract:
BDI agents have been used with considerable success for human modelling applications,
despite the fact that the framework was not designed specifically for this
purpose. I argue that the reason for BDI's success in this area is its folk
psychological roots, and go on to present an approach to systematically extending
the BDI framework without sacrificing the folk psychological underpinning.
I present two examples to demonstrate this approach. The first starts to
address the issue of human decision making, explicitly using folk psychological
concepts. The second tackles the issue of perception and action, and "bolts
on" to the existing framework, constraining its behaviour without directly
referring to the folk psychological constructs. These two examples demonstrate
the basis of the approach: to use folk psychological explanations where they
exist, and to constrain the reasoning without impacting on the core constructs
otherwise.
Biosketch:
Emma Norling is completing her PhD studies in the Department of Computer
Science and Software Engineering at the University of Melbourne. She has
been looking at BDI-based human modelling. Emma has a long history with
agents, being one of the beta-testers for dMARS in 1993. She was
coorganiser of the Melbourne Agent Systems School in 2003, and co-chair
for the Multi Agent Based Simulation (MABS) workshop held in conjunction
with AAMAS'03.