Experimental Reservoir Computing

March 2012

Reservoir computing aims to use recurrent dynamical systems for information processing. Our first publications reporting state of the art performance of analog reservoir computers have recently appeared in Nature Communications and Scientific Reports.

Read here the News and Views from Nature Physics on our work.

 Reservoir Computing aims to use recurrent dynamical systems for information processing.

Understanding how physical systems can process information is a major scientific challenge. The tremendous progress that has been accomplished during the past century has given rise to whole new fields of science, such as computer science and digital (silicon) electronics, or quantum information. Information processing by classical analog systems is a case apart because remarkable examples are found in the biological sciences (e.g. the brain), but our understanding is still very incomplete. Thus, we understand at the basic level many aspects of how information is processed using biochemical reactions in cells, or how information is processed by neurons in a brain, to take two examples. But how these elementary processes are organized at a higher level, and what is the large scale architecture of these systems, still escapes us. Understanding these issues would be of great conceptual interest. It could also have huge technological repercussions, as it could completely change the way we build information processing machines. That this tremendous scope for progress exists is illustrated by the approximately 6 orders of magnitude gap in energy consumption between a brain and a present day silicon computer.

So far most work on information processing in analog systems has been based on imitating biological systems. This has given rise to the field of artificial neural networks. Reservoir computing is a novel line of attack which uses recurrent dynamical systems (i.e. systems with internal feedback loops) to process time dependent data. Despite a relatively recent history -the first papers on the topic date from 2002-, the performance of reservoir computing is comparable to, and sometimes even exceed, that of other approaches to machine learning one specific tasks such as speech recognition.

Reservoir computing suggests many promising new avenues to implement computation in analog dynamic systems. The requirements for reservoir computing to be computationally universal (in the analog sense) are very loose: the reservoir is required to have fading memory, to be excitable by the input, and a high dimensional readout must be possible. Many physical systems could be conceived that adhere to these rules. However turning these general ideas into a working machine in more diffcult. Our work shows how this can be achieved.

In 2010 we proposed a simple scheme for the experimental realisation of reservoir computing which uses a single non linear node and a delay line:
Reservoir Computing: a Photonic Neural Network for Information Processing
Y. Paquot, B. Schrauwen, J. Dambre, M. Haelterman, S. Massar
In Proc. of SPIE Vol. 7728 77280B-1 (2010)

In 2011 we co-authored the first experimental realisation of reservoir computing with performance comparable to state of the art digital implementations:
Information processing using a single dynamical node as complex system

L. Appeltant, M.C. Soriano, G. Van der Sande, J. Danckaert, S. Massar, J. Dambre,B. Schrauwen, C.R. Mirasso, I. Fischer
Nature Communication 2, 468 (2011)

In 2012 we reported state of the art performance in an opto-electronic experiment:
Optoelectronic Reservoir Computing

Y. Paquot, F. Duport, A. Smerieri, J. Dambre, B. Schrauwen, M. Haelterman, S. Massar
Scientific Reports 2, 287 (2012)

A similar experiment was reported in
Photonic information processing beyond Turing: an optoelectronic implementation of reservoir computing
L. Larger, M. C. Soriano, D. Brunner, L. Appeltant, J. M. Gutierrez, L. Pesquera, C. R. Mirasso, and I. Fischer
Optics Express, Vol. 20, Issue 3, pp. 3241-3249 (2012)

For a pedagogic introduction to experimental reservoir computing, see the supplementary material of the 2012 Sci. Rept article.