Institute for Theoretical Information Technology

Bio-Inspired Information Processing

Biological communication and information systems have evolved over millions of years. Although they have been optimized under different design criteria than recent man-made technical communication systems, both are subject to the same information theoretic principles.

We want to design channel models to describe the information flow and information processing by neural networks. Massive parallelism, quantization, and information fusion are principles which have to be included in such models, as inspired by the biological ideal.

Our research has two main objectives:

• Provide models to numerically simulate certain aspects of neuronal communication to assist biological research.
• Isolate useful features of biological communication systems to analyze their applicability for technical scenarios.

Related research topics and publications

• R. Mathar, A. Schmeink, Cooperative Detection over Multiple Parallel Channels: a Principle Inspired by Nature, Proceedings: IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), Toronto, Canada, September 2011.
• R. Mathar, A. Schmeink, A Bio-Inspired Approach to Condensing Information, Proceedings: IEEE International Symposium on Information Theory (ISIT), Saint-Petersburg, August 2011.
• M. Arts, S. Corroy, M. Gorin, M. Spehr, A. Schmeink, R. Mathar, Modelling Biological Systems using a Parallel Quantized MIMO Channel, Proceedings: The Tenth International Symposium on Wireless Communication Systems (ISWCS 2013), Ilmenau, Germany, August 2013.
• M. Gorin, C. Tsitoura, A. Kahan, K. Watznauer, D. R. Drose, M. Arts, R. Mathar, S. O'Connor, I. L. Hanganu-Opatz, Y. Ben-Shaul, M. Spehr, Interdependent Conductances Drive Infraslow Intrinsic Rhythmogenesis in a Subset of Accessory Olfactory Bulb Projection Neurons, The Journal of Neuroscience, vol. 36, no. 11, pp. 3127-3144, March 2016.
• M. Arts, M. Cordts, M. Gorin, M. Spehr, R. Mathar, A Discontinuous Neural Network for Non-Negative Sparse Approximation, arXiv:1603.06353 [cs.NE], March 2016

Related projects

• ITOLF - An Information Theoretic Approach to Stimulus Processing in the Olfactory System. This project aims at applying the above mentioned methodology specifically to model the olfactory system of mice. We are working on this project in collaboration with Prof. Dr. Marc Spehr from the Department of Chemosensation.

Related student work

Contact

Martijn Arts