U.S. Department of Energy (DOE), Retired, July 2005 (35 years)
Office: Naval Research Laboratory, Code 5515
Naval Center for Applied Research in Artificial Intelligence
Office phone: (202) 767-3349
e-mail: Bill.Kennedy@NRL.Navy.mil
or Kennedy@MLLab.com Check e-mail here
National Research Council Associate
(post-doc) at the Naval Research Laboratory.
Associate External Research Professor,
Krasnow Institute for Advanced Study,
Cognitive Science and Artificial Intelligence: the nature and characteristics of long-term, symbolic learning, using Soar, ACT-R, and applied to cognitive robotics and the human-robot interactions.
The Nature of Long-Term, Symbolic Learning. The goal of this basic research is to understand the nature and characteristics of long-term, symbolic learning and create a computational cognitive theory of long-term symbolic learning that does not suffer from the Utility Problem.
Expand NRL’s Work in Cognitive
Robotics and Human-Robot Interactions (HRI):
Cognitively-plausible Spatial Reasoning. The goal of this research is to develop a computational model of spatial reasoning that is cognitively-plausible. We are extending the ACT-R cognitive architecture by adding a spatial reasoning module that uses a symbolic spatial representation including: a 2D, cognitive map; tracking of a target’s movement; and projection of target’s future motion. We have integrated this spatial representation and reasoning with multi-modal HRI (speech and gesture) on StealthBot, a robot that attempts to covertly approach a moving target while working with a teammate.
Simulating Teammates, the Robot’s Shared Mental Model (SMM), and a Robotic Theory of Mind (TOM). Develop a computational theory of teamwork; add the ability of a robot to simulate and have a TOM for its teammates; and maintain a situational awareness of a robot’s teammates including a SMM that supports what the robot expects a teammate would do in a situation based on what the robot itself would do under those circumstances. TeamBot has these capabilities and will be compared to human performance data.
Recent Publications
Artificial Intelligence or Cognitive Science:
Kennedy, W.G., Bugajska,
M.D.,
Kennedy, W.G. and Trafton, J.G. (2007) Long-term Symbolic Learning, Cognitive Systems Research 8(3), September 2007 (pp 237-247). Elsevier.
Kennedy, W.G., Bugajska, M.D.,
Marge, M., Fransen, B.R., Adams, W., Perzanowski, D., Schultz, A.C., and Trafton,
J.G. (2007) Spatial
Representation and Reasoning for Human-Robot Interaction, Proceedings of the Twenty-Second Conference
on Artificial Intelligence (AAAI 2007) (pp
1554-1559).
Kennedy, W.G. and Trafton, J.G.
(2007) Using
Simulations to Model Shared Mental Models. Proceedings of the Eighth International
Conference on Cognitive Modeling
(pp 253-254).
Kennedy, W.G. and Trafton, J.G.
(2006) Long-term Learning in
Soar and ACT-R, Proceedings of the Seventh
International Conference on Cognitive Modeling (pp 162-168).
Kennedy, W.G. and De Jong, K.A.
(2003) Characteristics
of Long-Term Learning in Soar and its Application to the Utility Problem, Proceedings
of the Twentieth
International Conference on Machine Learning (pp 337-344).
Government or Performance-Based
Management:
Kennedy, W.G. (2007) “Energy, the Big Picture”, presentation for NRL’s Postdoctoral Consortium, 20 June 2007. (also available with speaker’s notes here)
Kennedy, W.G. (2006) Policy as a Separate Level for
Performance Management, presentation at The Atlanta Conference on
Science and Technology Policy 2006,
Kennedy, W.G. (2005) Different Levels of Performance Measures for
Different Uses (The “PPP Proposal”), presentation at Evaluation 2005: Crossing Boarders,
Crossing Boundaries, Joint Canadian Evaluation Society/American Evaluation
Association Conference, Toronto, Canada, October 26-29, 2005.
Links:
Curriculum Vitae (CV)
___________________
“My Little Lab”
www.MLLab.com
Updated 16 April 2008