Connectionists: speciel sessions at ESANN'2006 European Symposium on Artificial Neural Networks
esann
esann at dice.ucl.ac.be
Wed Oct 5 06:34:21 EDT 2005
ESANN'2006
14th European Symposium
on Artificial Neural Networks
14th European Symposium on Artificial Neural Networks
Advances in Computational Intelligence and Learning
Bruges (Belgium) - April 26-27-28, 2006
Special sessions
=====================================================
The following message contains a summary of all special sessions that will
be organized during the ESANN'2006 conference. Authors are invited to
submit their contributions to one of these sessions or to a regular session,
according to the guidelines found on the web pages of the conference
http://www.dice.ucl.ac.be/esann/.
According to our policy to reduce the number of unsolicited e-mails, we
gathered all special session descriptions in a single message, and try to
avoid sending it to overlapping distribution lists. We apologize if you
receive multiple copies of this e-mail despite our precautions.
Special sessions that will be organized during the ESANN'2006 conference
========================================================================
1. Semi-blind approaches for Blind Source Separation (BSS) and
Independent Component Analysis (ICA)
M. Babaie-Zadeh, Sharif Univ. Tech. (Iran),
C. Jutten, CNRS – Univ. J. Fourier – INPG (France)
2. Visualization methods for data mining
F. Rossi, INRIA Rocquencourt (France)
3. Neural Networks and Machine Learning in Bioinformatics - Theory
and Applications
B. Hammer, Clausthal Univ. Tech. (Germany),
S. Kaski, Helsinki Univ. Tech. (Finland),
U. Seiffert, IPK Gatersleben (Germany),
T. Villmann, Univ. Leipzig (Germany)
4. Online Learning in Cognitive Robotics
J.J. Steil, Univ. Bielefeld,
H. Wersing, Honda Research Institute Europe (Germany)
5. Man-Machine-Interfaces - Processing of nervous signals
M. Bogdan, Univ. Tübingen (Germany)
6. Nonlinear dynamics
N. Crook, T. olde Scheper, Oxford Brookes University (UK)
Short descriptions
==================
Semi-blind approaches for Blind Source Separation (BSS) and Independent
Component Analysis (ICA)
-----------------------------------------------------------------------
Organized by:
- M. Babaie-Zadeh, Sharif Univ. Tech. (Iran)
- C. Jutten, CNRS – Univ. J. Fourier – INPG (France)
In the original formulation of the Blind Source Separation (BSS) problem, it
is usually assumed that there is no prior information about source signals
but their statistical independence. The methods then try to separate the
sources by transforming the observations into as statistically independent
as possible outputs (ICA). A well known result states that decorrelation
(second-order independence) of the outputs is not sufficient for separating
the sources. Another well-known result states that separating Gaussian
sources using this approach is not possible.
However, simple prior information about source signals can lead to new
methods, whose simplicity and separation quality may significantly be
improved (in terms of samples size, algorithm simplicity and speed, ability
to separate a larger class of signals, etc.). For example, if we already
know that the source signals are temporally correlated, it is possible to
separate them by using second-order approaches : the algorithm based on
second-order statistics is then simpler, Gaussian sources can be separated.
We call such an approach a “Semi-Blind” approach, because although it is not
completely blind, the available prior information about the sources is very
weak and remain true for a large class of sources.
Some of the most famous priors for designing Semi-Blind approaches for BSS
are:
- Sparsity of the source signals: Such a prior enables us to separate more
sources than sensors, and even dropping the independence assumption. Hence,
these approaches are usually called Sparse Component Analysis (SCA).
- Temporal correlation of the source signals (colored sources) enables
separation of Gaussian sources, using second-order approaches.
- Non-stationarity of the source signals enables separation of Gaussian
sources, using second-order approaches.
- Bounded sources enables, for example, simple geometric approaches to be
used.
- Models for source distribution (Markovian, etc.) can reduce the solution
indeterminacies and improve separation performance.
- Bayesian methods is a general framework for handling priors.
Of course, mixture of priors, currently not very usual, could also be
exploited and provide new algorithms.
In this special session, we invite authors to submit papers illustrating the
use of the above priors in BSS and ICA contexts.
Visualization methods for data mining
-------------------------------------
Organized by:
- F. Rossi, INRIA Rocquencourt (France)
In many situations, manual data exploration remains a mandatory preliminary
step that provides insights on the studied problem and helps solving it. It
is also very important for reporting results of data mining tools in an
exploitable way. While statistical summaries and simple linear methods give
a some rough analysis of a data set, sophisticated visualization methods
allow human experts to discover information in an easier and more intuitive
way. A very successful example of neural based visualization tool is given
by Kohonen's Self Organizing Maps used together with component planes,
U-matrix, P-matrix, etc.
This session aims at bringing together researchers interested in
visualization methods both used as exploratory tools (before other data
mining methods) and as explanatory tools (after other data mining methods).
Submissions are encouraged within (but not restricted to) following areas:
- non linear projection
- graph based visualization
- cluster visualization
- visualization method for supervised problems
- visualization of non vector data
Neural Networks and Machine Learning in Bioinformatics - Theory and
Applications
-------------------------------------------------------------------
Organized by:
- B. Hammer, Clausthal Univ. Tech. (Germany)
- S. Kaski, Helsinki Univ. Tech. (Finland)
- U. Seiffert, IPK Gatersleben (Germany)
- T. Villmann, Univ. Leipzig (Germany)
Bioinformatics is a promising and innovative research field. Despite of a
high number of techniques specifically dedicated to bioinformatic problems
as well as successful applications, we are in the beginning of a process to
massively integrate the aspects and experiences in the different core
subjects such as biology, medicine, computer science, engineering,
chemistry, physics and mathematics. Within this rather wide area we focus on
neural networks and machine learning related approaches in bioinformatics
with particular emphasis on integrative research in the background of the
above mentioned scope.
According to the high level and the aim of the hosting ESANN conference we
encourage authors to submit papers containing
- New theoretical aspects
- New methodologies
- Innovative applications
in the field of bioinformatics. A prospective but nonexclusive list of
topics is
- Genomic Profiling
- Pathways
- Sequence analysis
- Structured data
- Time series analysis
- Context related metrics in modelling and analysis
- Visualization
- Pattern recognition
- Image processing
- Clustering and Classification
- ...
Online Learning in Cognitive Robotics
-------------------------------------
Organized by:
- J.J. Steil, Univ. Bielefeld
- H. Wersing, Honda Research Institute Europe (Germany)
In hard- and software we currently observe technological breakthroughs
towards cognitive agents, which will soon incorporate a mixture of
miniaturized sensors, cameras, multi-DOF robots, and large data storage,
together with sophisticated artificial cognitive functions. Such
technologies might culminate in the widespread application of humanoid
robots for entertainment and house-care, in health-care assistant systems,
or advanced human-computer interfaces for multi-modal navigation in
high-dimensional data spaces. Making such technologies easily accessible for
every day use is essential for their acceptance by users and customers. At
all levels for such systems learning will be an essential ingredient to meet
the challenges in engineering, system development, and system integration
and neural network methods are of crucial importance in this arena.
Cognitive robots are meant to behave in the real world and to interact
smoothly with their users and the environment. While off-line learning is
well established to implement basis modules of such systems and many
learning methods work well in toy domains, in concrete scenarios on-line
adaptivity is necessary in many respects: in order to cope with the
inevitable uncertainties of the real world, the limited predictability of
the interaction structure, to acquire new and enhance preprogrammed
behavior. Online-learning is also the main methodological ingredient in the
developmental approach to intelligent robotics, which aims at incremental
progressing from simple to more and more complex behavior.
The current session will focus exclusively on the more difficult field of
online learning in real systems with real data. Given their systems meet
these constraints, authors are invited to submit contributions for all kinds
of cognitive robotics, for instance
- cognitive vision (eg. visual object learning, acquisition of visual
memory, adaptive scene analysis)
- localization and map building in mobile robots
- online trajectory learning and acquisition
- adaptive control of multi-DOF robots
- learning in behavioral architectures
- learning by demonstration and imitation
Man-Machine-Interfaces - Processing of nervous signals
------------------------------------------------------
Organized by:
- M. Bogdan, Univ. Tübingen (Germany)
Recently, Man-Machine-Interfaces contacting the nervous system in order to
extract information resp. to introduce information gain more and more in
importance. In order to establish systems like neural prostheses or
Brain-Computer-Interfaces, powerful (real time) algorithms for processing
nerve signals or their field potentials are requested. Another important
point is the introduction of informations into nervous systems be means like
functional electrical stimulation (FES).
Topics of this session can be, but are not limited to NN-based algorithms
and applications for
- Neural Prostheses
- Brain-Computer-Interfaces
- Multi Neuron Recordings
- Multi Electrode Arrays
- Functional Electrical Stimulation
- Population Coding
- Spike Sorting
- ...
Nonlinear dynamics
------------------
Organized by:
- N. Crook, T. olde Scheper, Oxford Brookes University (UK)
The field of nonlinear dynamics has been a useful ally in the study of
artificial neural networks (ANNs) in recent years. Investigations into the
stability of recurrent networks, for example, have helped to define the
characteristics of weight matrixes which guarantee stable solutions. Similar
studies have assessed the stability of Hopfield networks with distributed
delays. However, some have suggested that nonlinear dynamics should play a
more central role in models of neural information processing. Observations
of the presence of chaotic dynamics in the firing patterns of natural
neuronal systems has added some support to this suggestion. A range of
models have been proposed in the literature that place nonlinear dynamics at
the heart of neural information processing. Some of these use chaos as a
basic for neural itinerancy, a process involving deterministic search
through memory states. Others use the bifurcating properties of specific
chaotic systems as a means of switching between states. This session will
open with a tutorial paper outlining these different approaches. The session
will also include paper contributions by some leading authors in the field.
========================================================
ESANN - European Symposium on Artificial Neural Networks
http://www.dice.ucl.ac.be/esann
* For submissions of papers, reviews,...
Michel Verleysen
Univ. Cath. de Louvain - Microelectronics Laboratory
3, pl. du Levant - B-1348 Louvain-la-Neuve - Belgium
tel: +32 10 47 25 51 - fax: + 32 10 47 25 98
mailto:esann at dice.ucl.ac.be
* Conference secretariat
d-side conference services
24 av. L. Mommaerts - B-1140 Evere - Belgium
tel: + 32 2 730 06 11 - fax: + 32 2 730 06 00
mailto:esann at dice.ucl.ac.be
========================================================
More information about the Connectionists
mailing list