Second
Workshop On
Self-Organization In Representations
For Evolutionary Algorithms:
Building complexity from simplicity
To be held as part of the
Genetic
and Evolutionary Computation Conference (GECCO-2005)
June 25-29, 2005 (Saturday - Wednesday)
Loew's L'Enfant Plaza Hotel
Washington, DC, USA
www.isgec.org/GECCO-2005
Organized by ACM SIG-EVO
|
Important Dates
Submission Deadline:
14 April, 2005
Author Notification:
24 April, 2005
Camera-Ready Deadline:
27 April, 2005
Chairs
Ivan I. Garibay*
Sanjeev Kumar**
Ozlem Garibay*
Hal Stringer*
*Evolutionary Computation
Laboratory, University of Central Florida
**Krasnow Institute
for Advanced Studies, George Mason University
Program Committe
Josh Bongard, Sibley School of Mech. and Aerospace Eng.,
Cornell University
Peter Eggenberger, Dept. of Computer Science, University
of Zurich
Gregory Hornby, Computational Sciences Div. at NASA Ames
Research Center
Ivan Garibay, School of Computer Science, University
of Central Florida
Ozlem Garibay, School of Computer Science, University
of Central Florida
Sanjeev Kumar, Krasnow Institute for Advanced Study,
George Mason University
W. B. Langdon Computer Science, University College, London
Joseph Lewis, Computer Science Department, San Diego
State University
Julian Miller, Department of Electronics, University
of York
Tim Otter, Department of Biology, Albertson College of
Idaho
Paul Wiegand, Naval Research Labs. Washington D.C.
Annie Wu, School of Computer Science, University of Central
Florida
Contact Info
Washington, June 26, 2005 (8:30AM-12:30PM)
The success of Evolutionary Algorithms (EAs) on a wide range of otherwise intractable problems has promoted its use. As EAs are applied to increasingly difficult problems that require increasingly complex solutions, they face a number of problems: premature convergence to suboptimal solutions, stagnation of search in large search spaces, negative epistatic effects, disruption of large building blocks, and scalability, among others. Nature evolves instructions in the form of genes that are used to specify the construction of organisms using a highly non-linear process: development. Self-organization is fundamental to the developmental process at all levels: molecular, genetic, and cellular. With new reports of the number of genes in the human genome being revised downwards, the role of self-organization in complex webs of gene regulation is all the more salient. Given these new findings, perhaps the self-organization of genotypic instructions that transform genotype to phenotype is a key missing ingredient necessary for unleashing the evolution of complex and scalable solutions with emergent phenomena such as: scale-free-ness, adaptability, innovation, evolvability, and robustness. This workshop will focus on domain-independent methods for representing complex solutions with self-organizable building blocks, and on developmental principles for specifying the construction of complex systems. The workshop welcomes submissions from biologists on relevant biology that may help shed more light on self-organizing principles for evolutionary computation.Topics of interest include
- Models of complexity building using self-organization
- Emergent behavior in representations
- Methods of design and evaluation of self-organizable representational building blocks
- Scalability of self-organizational processes to high complexities
- Self-organization theoretical approaches: complexity, chaos, synergetics, self-organized criticality, non-equilibrium thermodynamics, etc.
- Self-organized development
- Genotype-phenotype mappings for self-organization and single & multicellular development
- Pattern formation, morphogenesis, cellular differentiation, and growth
- Models of genetic regulatory networks, modularity, segmentation, and compartmentalization
- Scalability & Evolvability of developmental processes
- Robustness, self-repair and regeneration in developmental processes
- Real world applications of developmental principles
Proceedings Workshop Self-organization in Representations for Evolutionary Algorithms
Preface
Ivan Garibay, Sanjeev Kumar, Ozlem Garibay, and Hal Stringer.
Invited Speaker Talk
Dr. Julian Miller
On location independent representations and self-organization
Ivan Garibay and Annie S. Wu and Ozlem Garibay
[PDF Full-Text]
Method Trees: Building Blocks for Self-Organizable Representations of Value Series
Ingo Mierswa and Katharina Morik
[PDF Full-Text]
Genotype, Phenotype and Ontogeny
Tim Otter
[PDF Full-Text]
Behaviorally Coupled Emergent Representation
Joseph Lewis and Jamie Lawson
[PDF Full-Text]
Self-Organizing Networks described by Evolutionary Algorithms
Katharina A. Lehmann
[PDF Full-Text]
A Developmental Genetics-Inspired Approach to Robot Control
Sanjeev Kumar
[PDF Full-Text]
Workshop Presentation Schedule
8:30am-8:35am Introduction by Organizers
8:35am-9:10am Invited speaker: Dr. Julian Miller, University of York
9:10am-9:35am I. Garibay, A. Wu, O. Garibay"On location independent representations and self-organization"
9:35am-10:00am I. Mierswa, K. Morik "Method Trees: Building Blocks for Self-Organizable Representations of Value Series"
10:00am-10:25am T. Otter, "Implementation of a Prototypal Simulator
for a Formal Model Based on Pattern Matching and Recombination"
10:25am-10:45am BREAK
10:45am-11:10am J. Lewis, J. Lawson "Behaviorally Coupled Emergent Representation"
11:10am-11:35am K. Lehmann, "Self-Organizing Networks described by Evolutionary Algorithms"
11:35am-12:00pm S. Kumar, "A Developmental Genetics-Inspired Approach to Robot Control"
12:00pm-12:30pm Panel discussion
Workshop Format
This workshop seeks to bring together researchers from diverse problem domains to informally discuss issues related to the representation of complex solutions using self-organization of simple building blocks for evolutionary algorithms in particular, and the issue of building complexity from simplicity in general. We plan to have a series of short technical presentations followed by a panel discussion.
We welcome technical papers describing completed or on-going research as well as position papers outlining current research issues, approaches or research agendas. We also welcome suggestions to panel discussions.
Workshop Submission Instructions
Please submit proposed contributions via email to igaribay@cs.ucf.edu in PS or PDF format by March 14th. Accepted contributions will be published in the GECCO-2005 CD-ROM.
Attendance
Open to all GECCO 2005 attendeesMore information
For more information, comments or suggestions please email Ivan Garibay at igaribay@cs.ucf.eduRelated Links