EvoWorkshops2006: EvoPhD

1st European Graduate Student Workshop on Evolutionary Computation

This is the first European workshop on evolutionary computation that focuses on work of PhD students. Its main aim is to give students feedback on the current state of their thesis. This workshop provides a chance to students to present their work to a friendly audience of other students as well as experts in the field. It also provides students with contacts and professional networking opportunities, which helps them to integrate into the community.

The proceedings of the 2006 workshop can be downloaded in PDF format: EvoPhD2006 Proceedings.

For you as a student, the aim of a submission is to get feedback on the current state of your thesis. Submissions will be evaluated by members of a high-quality committee that consists of long running members of the community with the goal to provide in-depth feedback. You are required to submit a paper that summarizes the research performed and planned as part of your dissertation.

The number of participants is limited. Accepted papers will not be published in a proceedings, but will be disseminated among the participants of the event.

The workshop is part of EuroGP2006 & EvoCOP2006, incorporating EvoWorkshops2006, which combined form Europe's premier co-located events in the field of evolutionary computing. The next event takes place in Budapest, Hungary. Featuring the latest in theoretical and applied research, the topics include recent genetic programming challenges, evolutionary and other meta-heuristic approaches for combinatorial optimization, evolutionary algorithms in the biosciences, in music and art domains, in image analysis and signal processing systems, in hardware optimisation and as applied to a range of industrial and financial optimisation problems.

Web address: http://www.evonet.info/eurogp2006

Topics include, but are not limited to,
  • Any topic that fits in the scope of the conferences and workshops
  • Bio-inspired computing, such as evolutionary computing, ant colony optimization, swarm intelligence, and neural networks
  • Local optimisation methods (tabu search, simulated annealing)
  • All flavours of evolutionary computation (genetic programming, evolution strategies, genetic algorithms, memetic algorithms, etc.)
  • Combinatorial optimisation with bio-inspired computing
  • Application of evolutionary computation to real-life problems
  • Hybrid architectures that include bio-inspired components
  • Theory on a relevant area of bio-inspired computing

Accepted papers: titles and abstracts

An Evolutionary Approach for Neural Model Optimization
Antonia Azzini

Neuro-genetic systems are biologycally inspired computational models that use evolutionary algorithms (EAs) in conjunction with neural networks (NNs) to solve problems. The PhD research work presented in this paper describes an evolutionary approach to the joint optimization of neural network structure and weights which can take advantage of backpropagation as a specialized decoder. The approach is successfully applied to a real-world engine fault diagnosis problem and to a classification application of brain waves in the context of brain-computer interfaces.


Evolution of Small-World Networks as a support for Cellular Automata Computation
Christian Darabos
(Nominated for Best Paper Award)

We study an extension of cellular automata to arbitrary interconnection topologies for the majority and the synchronization problems. By using an evolutionary algorithm, we show that small-world type network topologies consistently evolve from regular and random structures without being designed beforehand. These topologies have better performance than regular lattice structures and are easier to evolve, which could explain in part their ubiquity. Moreover, we show experimentally these graph topologies are much more robust in the face of random faults than lattice structures for these problems.


Extended Particle Swarm to Simulate Biology-Like Systems
Cecilia Di Chio
(Nominated for Best Paper Award)

Is it possible to simulate socio-biological behaviours using particle swarm systems? And if so, what should it be the best approach to use? These are the questions which I would like to answer with my research. Particle swarm systems have been originally developed to model social behaviours. My research will therefore follow the initial socio-biological metaphor underlying particle systems. The idea is to use a genetic programming approach to automatically evolve the particle swarm equations to model animal social behaviours. This research is intended to be a first example of application of genetic programming and particle swarm to simulate animal behaviours.


Geometric Unification of Evolutionary Algorithms
Alberto Moraglio

Evolutionary algorithms are only superficially different and can be unified within an axiomatic geometric framework by abstraction of the solution representation. This framework describes the evolutionary search in a representation-independent way, purely in geometric terms, paving the road to a general theory of evolutionary algorithms. It also leads to a principled design methodology for the crossover operator for any solution representation.


``Good'' Observers Enhance SGA Exploration
Christophe Phillemote

Most metaheuristics try to find a good balance between exploitation and exploration to achieve their goals. The exploration efficiency is highly dependent on the cardinality and ruggedness of the search space. A metaheuristic like the Simple Genetic Algorithm (SGA) can suffer a lot when traversing very large landscapes, especially deceptive ones. The approach proposed here improves the exploration of the SGA through the use of behavioural information of the SGA itself. Behavioural information on the SGA is obtained through a number of competitive processes which we refer to as ``observers''. The new metaheuristic we investigate, trains the observers for a specific time and then decides which of them is the most suitable to solve the whole problem. Concretely, a second evolutionary stage has been added to evolve observers for the SGA. These observers transform the cardinality and ruggedness of the search space through a simplification of the genotype. To test the proposed approach, we chose some difficult problems such as the Hierarchical IF-and-only-iF (HIFF). We obtained very good results, since we seriously improved the adaptive capacity of the SGA. Based on the current results, we are encouraged to continue in this way.


Evolving Exressive performance through simulating artificial agents' interaction
Qijun Zhang
(Nominated for Best Paper Award)

This PhD project focuses on the usage of Evolutionary Computation (EC) in expressive music performance research. We aim at building a computational model that co-evolves agent performers and agent listeners. Through these autonomous agents' interactions in the society, we are hoping to observe the emergence of shared repertoire of expressive music performance, and ideally this has some similarity with human performances. The work has been done is the first stage of this PhD project, in which we have implemented a system that uses Genetic Algorithm (GA) to evolve hierarchical time vs. amplitude matrices for music interpretation. The fitness for the GA is decided by how well the interpretation fits some rules associated with the piece's structural characteristics.


Submission Information (version 1)

You can download this information as a text file.

The deadline for submission is 16 December, 2005
Deadline extension: 23 December, 2005.

Authors will be notified of acceptance on 3 February 2006.

Camera-ready version of accepted paper due on 3 March 2006.

The page limit is 12 A4 pages in Springer LCNS format.

The filesize should not exceed 1 Mb.

The review process is not double-blind, please have your full name on the submission.

Submissions should be send in uncompressed PDF by e-mail to evophd2006@vanhemert.co.uk

Present a proof of your PhD studentship in the form of a letter from your department, signed by your supervisor/promoter. Send this letter to:

Mario Giacobini
INFORGE
Amphipôle, Unil-Sorge
Université de Lausanne
CH-1015 Dorigny
Switzerland

or fax it to +41 21 692 35 85

Use the following template for your submission:

============================== BEGIN TEMPLATE ============================== 

1. TITLE, full name (first and last), institution/affiliation, e-mail
address

2. ABSTRACT: maximum 250 words abstract of the topic and goals

3. INTRODUCTION OF THE RESEARCH AREA

3.1. This should include a concise description of the problem you are
tackling. What are important research questions in the area, and what are
people undertaking to answer these questions.

3.2. A very short overview of the most important literature and the most
important achievements that are relevant to your study.

4. YOUR RESEARCH AND STUDY

4.1 GOALS

What are the goals of your research. In other words, what research
questions are you trying to answer. What is the priority of these
questions.

4.1 CURRENT STATUS

Provide a short overview of what you have done so far and remember that
not all achievements need to be positive. If you are comparing your new
algorithm to others and have not yet found better results, then this is
still an achievement. Use the list of goals to show how far you are with
the individual questions.

4.2 FUTURE PLANNING

Provide a short plan of the work you will do until the end of your PhD
study. Which questions will you most likely be able to address. What kind
of experiments or proofs will you attempt to answer these questions? How
will this fit into the time frame of your PhD study.

If appropriate, tell us about your plans for when you have finished your
PhD studies. Will you stay an academic or do you prefer to enter a
commercial career. Perhaps even some hybrid between these two? Will you
apply for any scholarships, grants or fellowships? Would you like to go to
another country?

4.3 STUDY

At what stage are you in your PhD study? How many years is your study in
total? If there is some mechanism that tracks your progress, such as study
points, then explain this and tell us where you are at.

5. RESULTS

When your work is empirical, give detailed explanations of the results you
have achieved, and how you obtained them. Be clear on your methodology and
also take care in explaining how results should be interpreted. For
theoretical work, provide an explanation of the theorems and where they
are of importance. Explain your proofs or partial proofs so that it is
easier for reviewers to understand your line of thought. This is the more
scientific part, which can include material from previously written
material. We encourage you to use the following subsections when including
empirical results.

5.1 METHODOLOGY

Explanation of the algorithms/methods used. 

5.2 EXPERIMENT

Benchmark files used, and where to obtain these. Parameter settings of
algorithms. Problem generators. Number of independent runs used. Make sure
that anyone reading your description can repeat the experiments and obtain
equal or highly similar results.

5.3 RESULTS

The presentation of results heavily depends on your questions, but we want
to provide some guidelines. Always make sure figures and tables are
discussed in the text. Label graphs and tables so their contents can be
understood. Use readable fonts and clear graphics. Only show results when
they contribute to answering your questions. Give detailed explanation of
how results are obtained. Provide statistical verification of variation
and significance of your results (standard deviations, confidence
intervals, t-tests, Wilcoxon tests, etc.)

6. ACHIEVEMENTS

Show what you have achieved, similar to the conclusions in a paper.
Moreover, explain what you think you have achieved when your thesis is
finished. Last, provide any open issues that you feel are important but
you find hard or impossible to address.

7. FEEDBACK

Tell us what you hope to gain from this Doctoral Consortium. What kind of
feedback do you hope to get? Contemplate this carefully, as this section
is the key into getting the answers you want from the experts.

8. BIBLIOGRAPHICAL REFERENCES

Make sure you cite work where appropriate. Also include your own
publications, as you will surely reference your work in some of the
sections.

============================== END TEMPLATE ============================== 

Programme Committee

Enrique Alba
Anne Auger
Stefano Cagnoni
Mathieu CapcarrŤre
Pierre Collet
Ernesto Costa
Aniko Ekart
Jens Gottlieb
Steven Gustafson
Bill Langdon
Tom Lenaerts
Evelyne Lutton
JJ Merelo
Julian Miller
Daniele Radicioni
GŁnther Raidl
Peter Ross
Franz Rothlauf
Conor Ryan
Giovanni Squillero
Christine Solnon
Marco Tomassini
Leonardo Vanneschi
Sébastien Verel

Organising Committee

Program Chairs
Mario Giacobini
Mario.Giacobini AT unil DOT ch
University of Lausanne, Switzerland
 
Jano van Hemert
University of Edinburgh, UK
 
EvoWorkshops2006 Chair
Franz Rothlauf
rothlauf AT uni-mannheim DOT de
University of Mannheim, Germany
 
Local Chair
Anikó Ekárt
ekart AT sztaki DOT hu
Hungarian Academy of Sciences
 
Publicity Chair
Steven Gustafson
smg AT cs DOT nott DOT ac DOT uk
University of Nottingham, UK

Note: the e-mail addresses are masked for spam protection.