================================================================
Call
for Papers
International Workshop of the
21st Australasian Joint Conference on Artificial Intelligence (AI 08)
Ensemble Learning in Data Mining and Pattern Recognition
Auckland,
New Zealand, December 1-2,
2008
================================================================
In
ensemble learning
multiple models are combined under the assumption that two or even
more of them are better than only one. Decisions of multiple
hypotheses are fused in order to produce an overall hypothesis
combing to accurate and robust results. Various ensemble learning
algorithms have been developed during the last decade, and many of
them have been applied with great success to real world applications,
particularly in the field of pattern recognition, e.g. visual 3D
object recognition, face detection and recognition. In the context of
multiple models for pattern recognition the term multiple classifier
system (MCS) is typically used for such models, the most popular
approaches of MCS are boosting or bagging. More recently, methods
for the combination of models trained by unsupervised learning
procedures have been proposed, for instance to improve the
reliability and validity of the discovered cluster structures in
cluster analysis applications.
Another important data mining problem
arises in applications where the a priori defined learning task
changes over time, e.g. through additional incoming data (changes or
shifts in the data set collection procedure or newly sampled data),
additional features (changes in the feature extraction procedures),
or new classes (changes in the task that has to be solved). Examples
of such applications, are time series data, gene expressions,
partially-labeled data, noisy data and online applications where the
full data is not known in advance and must be collected and processed
in smaller subsets, e.g. sample by sample.
--------------------------------------------------------
Important
Dates
--------------------------------------------------------
Paper
due: 1 November 2008
Notification date: 7 November 2008
Camera
ready: 14 November 2008
Registration: 21 November 2008
Workshop:
1-2 December 2008
--------------------------------------------------------
Topics
of
Interest:
--------------------------------------------------------
We
welcome the following areas that are related to
Ensemble Learning
in Data Mining and Pattern Recognition
(but not limited to):
Supervised ensemble learning problems and multiple classifier systems
Theoretical aspects of classifier fusion and classifier selection
Semisupervised learning in multiple classifier systems
Unsupervised ensemble learning for unlabeled data
Information fusion and data fusion
Ensemble methods for regression and time series prediction
Supervised and unsupervised ensemble learning in changing environments
Application of ensemble learning in bioinformatics, medicine, web mining, multi-sensor-systems, multimodal man-machine-interfaces.
--------------------------------------------------------
Workshop
Organizer
--------------------------------------------------------
Friedhelm Schwenker (University of Ulm, Germany)
--------------------------------------------------------
Program
Committee
--------------------------------------------------------
Shigeno Abe (University of Kobe, Japan)
Monica Biancchi (University of Siena, Italy)
Nick Campbell (National Institute of Information and Communications Technologies and University of Kobe, Italy)
Neamat El Gayar (Cairo University, Egypt)
Simone Marinai (University of Florence, Italy)
Günther Palm (University of Ulm, Germany)
Lionel Prevost (Pierre et Marie Curie University Paris, France)
Edmondo Trentin (University of Siena, Italy)
--------------------------------------------------------
Submission
Requirements
--------------------------------------------------------
We
welcome papers presenting new original work. Submitted papers should
not exceed a length of 10 pages. We prefer Springer LNAI style (for
further information and style files see
http://www.springer.com/computer/lncs).
Papers will be reviewed on overall quality and relevance. All
accepted papers will be fully published in the workshop proceedings.
It is planned to publish extended versions of the accepted papers in
the Springer LNAI series or in a special issue of Pattern
Recognition Letters.
Submissions should be sent to the workshop organizer:
Friedhelm Schwenker (Email: friedhelm.schwenker@uni-ulm.de)
At
least one author of each accepted paper is required to attend the
workshop.
--------------------------------------------------------
Location
--------------------------------------------------------
This
workshop will be co-located with the 21st
Australasian Joint
Conference on Artificial Intelligence
(AI-08) which will be held
at Quay West Hotel, Auckland,
New Zealand on December 1-2, 2008.