Information processing using a single dynamical node as complex system

By L. Appeltant, Miguel C. Soriano, G. Van der Sande, J. Danckaert, Serge Massar, Joni Dambre, Benjamin Schrauwen, Claudio R. Mirasso, Ingo Fischer. Nature Communications, 2011

1,486 citations recorded by the source corpus.

Novel methods for information processing are highly desired in our information-driven society. Inspired by the brain's ability to process information, the recently introduced paradigm known as 'reservoir computing' shows that complex networks can efficiently perform computation. Here we introduce a novel architecture that reduces the usually required large number of elements to a single nonlinear node with delayed feedback. Through an electronic implementation, we experimentally and numerically demonstrate excellent performance in a speech recognition benchmark. Complementary numerical studies also show excellent performance for a time series prediction benchmark. These results prove that delay-dynamical systems, even in their simplest manifestation, can perform efficient information processing. This finding paves the way to feasible and resource-efficient technological implementations of reservoir computing.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/97e82bdbb17e0a79e21c54a504ddac2494dd2d66

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

In Electrical and Electronic Engineering

Citations in the corpus

Parallel photonic information processing at gigabyte per second data rates using transient states (2013, background): “7).”
Complex photonics: Dynamics and applications of delay-coupled semiconductors lasers (2013, methodology): “It was shown that a single nonlinear element with delay is capable of performing complex classification tasks by using its transient response when its own dynamics are mixed with an external information stream (Appeltant et al., 2011).”
Randomness in neural networks: an overview (2017, background): “Together, these results raise some important questions, namely: are the random ‘echoes’ produced by ESNs actually helpful, or can we discard randomness and focus on the algebraic properties of the matrices? Furthermore, in this case, is this true for any possible class of problems? Recently, Appelta”
Information processing via physical soft body (2015, background): “Lately, electronic and optical implementations have also been reported 20–23 .”
Reinforcement Learning in a large scale photonic Recurrent Neural Network (2017, background): “However, quickly it was realized that RC is highly attractive for analog hardware implementations [4,5].”
Continuous-variable quantum neural networks (2018, background): “[26] Lennert Appeltant, Miguel Cornelles Soriano, Guy Van der Sande, Jan Danckaert, Serge Massar, Joni Dambre, Benjamin Schrauwen, Claudio R Mirasso, and Ingo Fischer.”
Controlling Recurrent Neural Networks by Conceptors (2014, methodology): “Following the terminology of the reservoir computing [56, 6], I refer to such randomly connected neural networks as reservoirs.”
Deep physical neural networks trained with backpropagation (2021, methodology): “21, 22).”
Experimental demonstration of reservoir computing on a silicon photonics chip (2014, background): “Optical reservoirs based on a ﬁbre and one dynamical node 14–18 as well as reservoirs based on ring resonators 19 have been demonstrated.”