BioSPICE Developer page (private).
Tree-Augmented Naive Bayes (TAN) Classifier
Unfortunately at this time this project has no homepage. Look here in the future for more information about TAN and availability of TAN software.
This work was done under the DARPA High Performance Knowledge Base project. It has subsequently been used in the
CALO (registration required) project for
TOQR.
Previous Projects
Automating Exception Handling
See the
Air Mobility Command Execution Assistant Homepage
for further details.
Coordinating Distributed Activities (CODA)
See the CODA Homepage for details.
Our task was to develop a planning and decision aid (PDA) for DARPA's Small
Unit Operations (SUO) program, to show the feasibility of using advanced
planning technologies in SUO. The SUO concept envisions small units of
warfighters, separated and dispersed throughout a large battlespace, equipped
with personal equipment providing robust geolocation, computing, and
communications capabilities to allow units to effectively operate. The PDA
supports the generation of machine-understandable plans, and the
monitoring of their execution, using events as they are realistically
reported in the battlespace. Another goal of this project was to produce a
system that would successfully transition into the Phase III system of the
SUO prime contractor.
If commanders are flooded with more information than they can digest in a
timely manner, the resulting confusion may result in making worse decisions.
C2 software must understand the plan and situation in order to assist the
commander in translating information superiority into superior decisions.
We achieved our objective by showing the feasibility of using advanced
planning technologies in SUO. In the opinion of subject matter experts,
our representations modeled and monitored a SUO scenario
with enough fidelity to provide value to decision-makers. Our PDA used
a machine-understandable plan to generate appropriate alerts for the time and
location constraints in the plan, as well as for some more global constraints
such as fratricide risk. Alerts were generated in a timely manner without
overwhelming the user with too many alerts. Our execution monitoring
technology was able to easily monitor the volume of realistic data on the SUO
network (a dozen or more reports every second), while not being specific to
the plans used in our scenarios.
For further details, see our JAIR paper.
Air Campaign Planning Knowledge Base
The Air Campaign Planning Knowledge Base (ACP KB) is a knowledge base written for
the SIPE-2 generative planning system. It is used to
generate air campaign plans for the air superiority portion of an air campaign plan.
See the ACP Manual (PDF 123K)
for a description of the knowledge base.
The Air Campaign Planning Knowledge Base has been used in several projects,
including the Continuous Planning and Execution Framework (CPEF) and a
Technology Integration Experiment
for the
Multiagent Planning Architecture
Scientific Discovery for Molecular Biology Databases
Under this internally-funded project, we developed and applied
machine learning tools and techniques to databases that describe
physical and functional properties of proteins. The goal of the
project was to learn classifiers that will predict the function of an
enzyme from its structure. Our results were published in this
paper.
Publications
-
BioWarehouse: a bioinformatics database warehouse toolkit,
Thomas J. Lee, Yannick Pouliot, Valerie Wagner, Priyanka Gupta, David WJ Stringer-Calvert,
Jessica D Tenenbaum and Peter D Karp
BMC Bioinformatics 7:170 (23 March 2006).
-
Online Query Relaxation via Bayesian Causal Structures Discovery,
Ion Muslea and Thomas J. Lee,
Proceedings of the Twentieth National Conference on Artificial Intelligence (AAAI),
Pittsburgh, Pennsylvania, 2005.
-
Interactive Execution Monitoring of Agent Teams,
David E. Wilkins and Thomas J. Lee and Pauline Berry,
Journal of AI Research, volume 18, pages 217-261, March 2003.
-
Active Coordination
of Distributed Human Planners,
K. L. Myers, P. A. Jarvis, T. J. Lee,
Proceedings of the Sixth International Conference on AI Planning and
Scheduling (AIPS), 2002.
-
CODA: Coordinating Human Planners,
K. L. Myers, P. A. Jarvis, T. J. Lee,
Proceedings of the European Conference on Planning (ECP), Toledo, Spain, 2001.
-
Generating Qualitatively Different Plans through Metatheoretic Biases,
Karen L. Myers and Thomas J. Lee,
Proceedings of the Sixteenth National Conference on Artificial Intelligence (AAAI),
Orlando, Florida, 1999.
-
Bayesian Network Classification with Continuous Attributes: Getting the Best of Both Discretization and Parametric Fitting,
Nir Friedman, Moises Goldszmidt, and Thomas J. Lee,
Proceedings of the Fifteenth International Conference on Machine Learning (ICML),
Madison, Wisconsin, 1998.
-
Prediction of Enzyme Classification from Protein Sequence without the Use of Sequence Similarity,
Marie desJardins, Peter D. Karp, Markus Krummenacker, Thomas J. Lee, and Christos A. Ouzonis,
Proceedings of the Fifth International
Conference on Intelligent Systems for Molecular Biology (ISMB),
Halkidiki, Greece, 1997.
-
Using SIPE-2 to integrate planning for military air campaigns,
Thomas J. Lee and David E. Wilkins,
featured in IEEE Expert, December 1996,
as part of the Trends and Controversies section entitled
AI planning systems in the real world.
SRI Links:
Representation and Reasoning Program
||
AI Center
||
SRI International
Last modified: Wed Nov 02 14:06:48 PST 2005