Evolving Spiking Neural Networks: Growth of Learning Machines

J. David Schaffer gave an excellent talk on evolving Spiking Neural Networks at the Center for Collective Dynamics of Complex Systems (CoCo) Seminar Series.

Many of today's neural network applications are based on multi-layer implementations of the perceptron. A perceptron implements a neuron model that sums weighted inputs and applies a non-linear activation function to calculate the output. Despite that this model deviates from how biological neuron networks work, the approach works and is used until today. In contrast, Spiking Neural Networks (SNNs) are a type of neural network that increase the realism in neural network simulation by introducing a time aspect into the model. Other than perceptron networks that decide upon their output at each propagation cycle, SNNs fire when a certain membrane potential is reached, which puts information into the timing of a spike.

J. David Schaffer shows how a genetic algorithm can be applied to generate an SNN for a given problem. Therefore, the chromosome representing an SNN will be mapped onto a binary string which will be evolved with mutation and recombination.

Evolving Spiking Neural Networks: Growth of Learning Machines from Complex Systems on Vimeo

Further readings on the topic:

• Evolutionary Computation at Wikipedia
• Spiking Neural Network at Wikipedia 
• J. David Schaffer's Webpage at Binghampton University
• R. Batllori, Craig B. Laramee, W. Land, J. David Schaffer. Evolving spiking neural networks for robot control. In Procedia Computer Science (6):329-334, 2011. (fulltext)
• I. Fehervari and W. Elmenreich. Evolving neural network controllers for a team of self-organizing robots. Journal of Robotics, 2010. (fulltext)
• I. Fehervari and W. Elmenreich. Evolution as a tool to design self-organizing systems. In Self-Organizing Systems, volume LNCS 8221, pages 139–144. Springer Verlag, 2014. (fulltext)
• I. Fehérvári, A. Sobe, and W. Elmenreich. Biologically sound neural networks for embedded systems using OpenCL. In Proceedings of the International Conference on NETworked sYStems (NETYS 2013). Springer Verlag, May 2013. (fulltext)
• J. David Schaffer, Darrell Whitley, Larry J Eshelman. Combinations of genetic algorithms and neural networks: A survey of the state of the art. International Workshop on Combinations of Genetic Algorithms and Neural Networks (COGANN-92), pages 1-37, 1992.

Two Open PhD Student Positions at Alpen-Adria-Universität Klagenfurt, Austria

The Institute of Networked and Embedded Systems at the University of Klagenfurt, Austria, has an opening for two researchers (PhD student) in Self-Organizing Networked Systems.

You will work in an European project on Cyber-Physical Systems. The project duration is three years. Research will be conducted at the Smart Grids group at the Institute of Networked and Embedded Systems under the supervision of Professor Wilfried Elmenreich. Working language is English. The institute cooperates with national and international partners from industry and academia and is part of the research cluster Lakeside Labs (self-organizing networked systems).

Our team is very international and dedicated to quality research and teaching. The offices and laboratories are located in the well-equipped Lakeside Science & Technology Park. Working language is English. The Institute cooperates with national and international partners in research and industry. It is part of the research cluster Lakeside Labs on self-organizing networked systems.

The candidate is expected to have a master in electrical or computer engineering, computer science, applied mathematics or equivalent. Experience in one or more of the following fields is required:

• Distributed and swarm algorithms
• Genetic algorithms and evolutionary design
• Very good programming skills in Java

The contract will be based on 30 hours per week with a gross salary of at least 28314 EUR per year. Please mail applications containing a letter of interest, curriculum vitae, copies of academic certificates and courses, list of publications, and contact details of two references in a single PDF file to wilfried.elmenreich@aau.at before January 12, 2016. Women are especially encouraged to apply.

Tools for Calculating Academic Collaboration Distance

I think most of you have heard about the Erdös number. The Erdös number is the number of edges between you an Erdös in an author collaboration graph.
This is an undirected network where every published paper defines egdes between their authors. Having a low Erdös number somehow became a status symbol for researchers. Since Erdös already passed away, there is no way to get an Erdös number of one today, unless you hope for a Zombie apocalypse with the death rising:

Excerpt from "Apocalypse" by Randall Munroe at xkcd.com under CC-BY-NC 2.5

Due to Paul Erdös' outstanding publication productivity, there are quite a number of people with an Erdös number of 1, so if you find the right collaborator, you can reach an Erdös number of 2, if you like. But even beyond the fad on Erdös numbers, author collaboration graphs and distances between authors are an interesting way to define closeness between the work that two academics are doing.

What are good tools to calculate author collaboration distance?

There is MathSciNet, but their database only includes mathematical journals. Since my research is mostly published in computer science/embedded systems journals, this site doesn't work for me.
The zbMATH page offers a similar tool, again it seems to include only mathamatical journals. I should publish more there.

Previously, Microsoft Academic Research had a nice author collaboration search that graphically displayed the connections between any two authors. However, this feature is currently not available, since the page was restructured to work without the Silverlight plugin. I hope the feature comes back someday.

Distance calculator at csauthors.net

Currently, the best tool for computer scientists is the distance calculator at csauthors.net. It works with a database that seems to be more complete than the ones used by the sites cited above. The database is however far from being complete, so that distances are sometimes reported to be longer than they actually are.

My Erdös number

Thanks for asking! It is 3, for example via the following papers:

• W. Elmenreich, R. D'Souza, C. Bettstetter, and H. de Meer. A survey of models and design methods for self-organizing networked systems. In Proceedings of the Fourth International Workshop on Self-Organizing Systems, volume LNCS 5918, pages 37–49. Springer Verlag, 2009.
• D. Achlioptas, R. M. D'Souza and J. Spencer, "Explosive Percolation in Random Networks", Science, 323 (5920) 1453-1455, 2009.
• P. Erdös and J. Spencer. Imbalances in k-colorations. Networks, Volume 1, Issue 4, Pages 379–385, 1971 

All papers are on the topic of networks or networked systems. How fitting.