David W. Etherington
David Etherington is Director and Research Professor at the University
of Oregon's Computational Intelligence Research Laboratory CIRL. He is also Chief
Technology Officer of On Time
Systems, Inc., a startup company that is developing practical,
fielded, applications of CIRL's optimization technology in areas
ranging from aircraft routing to scheduling of ship construction.
He received his Ph.D. in Computer Science (Artificial Intelligence)
from the University of British Columbia in 1986. His dissertation was titled:
Reasoning with Incomplete Information: Investigations of
Non-monotonic Reasoning. From 1986 to 1993, when he left to help
start CIRL, he was a member of technical staff in the Artificial
Intelligence Principles Research department of AT&T Bell Laboratories,
in Murray Hill, NJ.
Dr. Etherington has published extensively, including a book on
nonmonotonic reasoning, which was published by Pitman/Morgan Kaufmann
in 1988. He was also founding chair of the International Federation
for Information Processing (IFIP) Working Group on Knowledge
Representation, has served as chair/co-chair/area-chair of several AI
workshops and conferences, and on the editorial boards of
Computational Intelligence (continuing) and the Journal of Artificial
Intelligence Research.
His current research interests include formal, tractable, theories of
knowledge representation, planning, optimization, and nonmonotonic
reasoning.
Research
- Context-based Nonmonotonic Reasoning
- To overcome the inherent intractability of nonmonotonic reasoning,
one can combine two known weak techniques -- limited contexts and
fast, incomplete consistency tests -- to obtain a powerful, tractable,
approximation mechanism. Typical subproblems include: formalizing the
conditions under which this approach gives justifiable results,
studying the mechanisms necessary for building and maintaining
contexts and for dealing with the errors induced by the approximation,
and implementing the ideas and evaluating them empirically against
non-trivial knowledge bases.
- Optimization
- CIRL has developed a variety of scheduling/optimization
techniques that are among the best in the world for their particular
areas of application. An interesting problem is expanding the
coverage of particular techniques, both to address more problems and
to make more techniques applicable to more problems, ideally allowing
greater flexibility and improved performance. Current projects
involve generalizing non-precedence-based scheduling techniques to
problems involving precedence constraints.
- Approximate Reasoning
- We have developed a general framework for approximating logical
deduction that admits the development of well-behaved, tractable
approximations. Typical subproblems include: formal analysis of the
general framework, developing and studying particular instances of the
framework, algorithm development and evaluation, and empirical
validation.
Teaching
In winter term, 1996, I taught the
Artificial
Intelligence Seminar (CIS607), which discussed tractable approaches to
logical deduction. You can see the
reading list.
Available Papers
A list of available papers can be found
here.
Contact Information
Address:
David W. Etherington
Computational Intelligence Research Laboratory
Mail:
CIRL
1269 University of Oregon
Eugene, OR 97403-1269
Courier:
CIRL, Ste 1
1850 Millrace Drive
Eugene, OR 97403
Phone:
+1-541-346-0470
Fax:
+1-541-346-0474
Email:
<ether <at> cirl.uoregon.edu>
PGP/GPG Public Key
My PGP/GPG public key can be found
here.
Back to CIRL's home page.