Optoelectronic Reservoir Computing

By Y. Paquot, François Duport, Anteo Smerieri, Joni Dambre, Benjamin Schrauwen, Marc Haelterman, Serge Massar. Scientific Reports, 2011

768 citations recorded by the source corpus.

Reservoir computing is a recently introduced, highly efficient bio-inspired approach for processing time dependent data. The basic scheme of reservoir computing consists of a non linear recurrent dynamical system coupled to a single input layer and a single output layer. Within these constraints many implementations are possible. Here we report an optoelectronic implementation of reservoir computing based on a recently proposed architecture consisting of a single non linear node and a delay line. Our implementation is sufficiently fast for real time information processing. We illustrate its performance on tasks of practical importance such as nonlinear channel equalization and speech recognition and obtain results comparable to state of the art digital implementations.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/9b1051e4466668ff4683cff7175561f8d43bf720

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 Electrical and Electronic Engineering

Citations in the corpus

High-Speed Photonic Reservoir Computing Using a Time-Delay-Based Architecture: Million Words per Second Classification (2017, background): “…step of the continuous dynamics [5] or, roughly, by neglecting the temporal nearest-neighbor coupling provided by the impulse response function (coupling was then induced through the hypothesis of an asynchronous configuration of the cyclic injected TDM samples with respect to the delay [7,8]).”
Broadcast and Weight: An Integrated Network For Scalable Photonic Spike Processing (2014, background): “For this reason, many attempts to leverage the capabilities of optics avoid a digital electronic computing paradigm altogether and instead target specialized tasks, including A/D conversion [41], amoeba-inspired processing in quantum dots [42], and reservoir computing [43], [44].”
Randomness in neural networks: an overview (2017, methodology): “Later, fully optoelectronic implementations of ESN schemes were independently presented in Paquot et al.123 and Larger et al.,124 paving the way to ultrafast implementations of RC using nonconventional computing technologies.”
Emerging opportunities and challenges for the future of reservoir computing (2024, methodology): “The target output are the corresponding labels which can be spoken digits 16,28,29,34,35,65,68,82 – 84 , image labels 33,35,85 – 87 , bit-symbols 16,29,68,88 – 93 and so on.”
Reservoir Computing Trends (2012, background): “Potential and functioning examples include analog electronics [40,39], randomly crystallized nonlinear electronic networks [44], opto-electronic [26,36] and optical [47] systems, or just a bucket filled with water [12].”
Large-Scale Optical Neural Networks based on Photoelectric Multiplication (2018, methodology): “Previous approaches used optoelectronics (photodiodes, lasers, amplifiers) both to sum neuron inputs [32, 24] and to generate nonlinearity or spiking dynamics [33, 34, 35, 36, 37]; here, thanks to the optical weight encoding, the synaptic weighting itself is performed opto-electronically.”
Integration of nanoscale memristor synapses in neuromorphic computing architectures (2013, background): “For example, reservoir computing (RC, also known under the terms Liquid State Machines and Echo State Networks) is a powerful general principle for computation and learning with complex dynamical systems such as recurrent networks of analog and spiking neurons [85, 86] or optoelectronic devices [87]”
Parallel photonic information processing at gigabyte per second data rates using transient states (2013, background): “Both tasks are standard to the machine learning community and allow for a direct comparison between different approaches to information processing 7,10,11,14 .”
Recent Advances in Physical Reservoir Computing: A Review (2018, background): “Spoken digit recognition (Verstraeten, Schrauwen, Stroobandt and Van Campenhout, 2005) Waveform classification (Paquot et al., 2012) Human action recognition (Soh & Demiris, 2012) Handwritten digit image recognition (Jalalvand, Van Wallendael, & Van de Walle, 2015) Time series forecasting Chaotic…”
Complex photonics: Dynamics and applications of delay-coupled semiconductors lasers (2013, methodology): “An analog computation concept based on a delayed-feedback system has been shown to perform certain tasks, such as spoken digit recognition and time series prediction, as well as other types of machine-learning approaches or even outperforming them (Appeltant et al., 2011; Duport et al., 2012; Larger”