With apologies for multiple copies:
* Deadline extended to April 17 *
The AAAI-06 Workshop on Learning for Search
http://www.cs.ubc.ca/~hutter/aaai06_ws
Purpose
=======================
Heuristic search is among the most widely used techniques in AI. In its
different varieties, such as tree-based search and local search, it
provides the core engine for applications as diverse as planning, parsing,
and protein folding. One of the most promising avenues for developing
improved search techniques is to use some kind of algorithmic component
that learns from experience. Many disparate techniques have arisen in
recent years that exploit learning to improve search and
problem-solving. These techniques can be off-line or on-line, based on hard
constraints or probabilistic biases, and applied to tree-structured or
local search.
This workshop aims to bring together researchers and practitioners from the
various subcommunities where such methods have arisen in order to learn
from each other, develop common understandings, and inspire new algorithms
and approaches.
Topics
======
Relevant topics include, but are not limited to:
* adaptive and self-tuning algorithms
* automated parameter tuning
* automated portfolio design
* clause learning
* computing search space features
* decision-theoretic approaches to learning in search
* dynamic portfolio design
* exploiting models of search spaces
* exploiting performance profiles or run-time distributions
* incremental and active learning in search
* learning to select operators or heuristic functions
* metareasoning from experience
* model-based search
* reinforcement learning for search algorithms
* runtime prediction
* speed-up learning
* uncertainty in runtime prediction
Please note that in addition to the "Learning for Search" workshop, there
will also be a separate workshop at AAAI-06 on "Heuristic Search,
Memory-based Heuristics and Their Applications". These two workshops will
be held on separate days and people are encouraged to submit to both
workshops if they wish.
Submission requirements
=======================
Potential participants wishing to present their work should submit
technical papers formatted in the AAAI conference style. Technical papers
are encouraged to be 6 pages in length with a maximum of 8 pages. Please
note that all submitted papers will be carefully peer-reviewed by multiple
reviewers for quality and relevance.
Other potential participants should submit a statement of relevant research
interests, maximum 2 pages in length. Note that all accepted submissions
will appear in the workshop notes.
Due to a request by AAAI to accept submissions until after the notification
date for the AAAI technical program, the general submission deadline for
this workshop was extended to April 17 (3 days after the notification
date).
Submissions should be sent in PDF format via email to
[EMAIL PROTECTED] Also note that per AAAI policy, participation
in the workshop is by invitation only and all workshop participants must
register for the main AAAI-06 conference.
Anyone interested in the workshop topic is invited to join the Yahoo group
http://groups.yahoo.com/group/learning_for_search/
Program Committee
===============
Wheeler Ruml (cochair), Palo Alto Research Center, USA
Frank Hutter (cochair), University of British Columbia, Canada
Tom Carchrae, Cork Constraint Computation Center, Ireland
Susan Epstein, City University of New York, USA
Youssef Hamadi, Microsoft Research Cambridge, UK
Henry Kautz, University of Washington, USA
Sven Koenig, University of Southern California, USA
Kevin Leyton-Brown, University of British Columbia, Canada
Toby Walsh, University of New South Wales, Australia
Shlomo Zilberstein, University of Massachusetts Amherst, USA
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
