Speech recognition is a critical task in the field of artificial intelligence (AI) and has witnessed remarkable advancements thanks to large and complex neural networks, whose training process typically requires massive amounts of labeled data and computationally intensive operations. An alternative paradigm, reservoir computing (RC), is energy efficient and is well adapted to implementation in physical substrates, but exhibits limitations in performance when compared with more resource-intensive machine learning algorithms. In this work, we address this challenge by investigating different architectures of interconnected reservoirs, all falling under the umbrella of deep RC (DRC). We propose a photonic-based deep reservoir computer and evaluate its effectiveness on different speech recognition tasks. We show specific design choices that aim to simplify the practical implementation of a reservoir computer while simultaneously achieving high-speed processing of high-dimensional audio signals. Overall, with the present work, we hope to help the advancement of low-power and high-performance neuromorphic hardware.

Reservoir computing is a bio-inspired computing paradigm with the main advantage of being less computationally demanding than state-of-the-art neural networks. However, the performance of reservoir computing is limited comparing to other power-hungry machine learning algorithms. Deep reservoir computing addresses this problem by using multiple interconnected reservoirs to provide greater capacity for processing information, leading to improved performance. In this work we propose a photonic deep reservoir computer: we test it on the speech recognition task, and we experimentally confirm its effectiveness. We show how some design choices can simplify the realization of a reservoir computer. Our photonic system is the first experimental deep reservoir computer that can process speech recordings in real-time. We thus hope that this work can pave the way for efficient dedicated hardware for high-speed signal processing.

Reservoir computing is a machine learning approach to designing artificial neural networks. Despite the significant simplification of the training process, the performance of such systems is comparable to other digital algorithms on a series of benchmark tasks. Recent investigations have demonstrated the possibility of performing long-horizon predictions of chaotic systems using reservoir computing. In this work we show that a trained reservoir computer can reproduce sufficiently well the properties a chaotic system, hence allowing full synchronisation. We illustrate this behaviour on the Mackey-Glass and Lorenz systems. Furthermore, we show that a reservoir computer can be used to crack chaos-based cryptographic protocols and illustrate this on two encryption schemes.

Reservoir Computing is a bio-inspired computing paradigm for processing time dependent signals (Jaeger andHaas 2004; Maass, Natschläger, and Markram 2002). It canbe easily implemented in hardware. The performance ofthese analogue devices matches digital algorithms on a series of benchmark tasks (see e.g. (Soriano et al. 2015) fora review). Their capacities could be extended by feedingthe output signal back into the reservoir, which would allow them to be applied to various signal generation tasks(Antonik et al. 2016b). In practice, this requires a high-speed readout layer for real-time output computation. Herewe achieve this by means of a field-programmable gate array (FPGA), and demonstrate the first photonic reservoircomputer with output feedback. We test our setup on theMackey-Glass chaotic time series generation task and obtain interesting prediction horizons, comparable to numerical simulations, with ample room for further improvement.Our work thus demonstrates the potential offered by the output feedback and opens a new area of novel applications forphotonic reservoir computing.

Reservoir Computing is a bio-inspired computing paradigm for processing time dependent signals. It can be easily implemented in hardware. The performance of these analogue devices matches digital algorithms on a series of benchmark tasks. Their capacities could be extended by feeding the output signal back into the reservoir, which would allow them to be applied to various signal generation tasks. In practice, this requires a high-speed readout layer for real-time output computation. Here we achieve this by means of an FPGA chip, and demonstrate the first photonic reservoir computer with output feedback. We test our setup on the Mackey-Glass chaotic time series generation task and obtain interesting prediction horizons, comparable to numerical simulations, with ample room for further improvement. Our work thus demonstrates the potential offered by the output feedback and opens a new area of novel applications for photonic reservoir computing.

Reservoir computing is a bioinspired computing paradigm for processing time-dependent signals. The performance of its analog implementation is comparable to other state-of-the-art algorithms for tasks such as speech recognition or chaotic time series prediction, but these are often constrained by the offline training methods commonly employed. Here, we investigated the online learning approach by training an optoelectronic reservoir computer using a simple gradient descent algorithm, programmed on a field-programmable gate array chip. Our system was applied to wireless communications, a quickly growing domain with an increasing demand for fast analog devices to equalize the nonlinear distorted channels. We report error rates up to two orders of magnitude lower than previous implementations on this task. We show that our system is particularly well suited for realistic channel equalization by testing it on a drifting and a switching channel and obtaining good performances.

Because of the small energy available aboard a satellite, the power amplifier must achieve a very high power efficiency which suggest to work close to the saturation point. This would be power efficient, but unfortunately would add non-linear distortions to the communication channel. Several equalization algorithms have been proposed to compensate for this non-linear behaviour. The Echo State Network (ESN), an algorithm coming from the field of artificial neural networks, has also been proposed for this task but has never been compared to state-of-the-art equalizers for non-linear channel. The aim of this paper is to adapt the ESN to the satellite communication channel and to compare it to the baseband Volterra equalizer. We show that the ESN is able to reach the same performances as the Volterra equalizer, evaluated in terms of bit error rate, and has similar complexity. In addition, we propose a new training strategy for the ESN and the Volterra equalizer to improve their performance.

We solve a 20-year old problem posed by Yannakakis and prove that no polynomial-size linear program (LP) exists whose associated polytope projects to the traveling salesman polytope, even if the LP is not required to be symmetric. Moreover, we prove that this holds also for the cut polytope and the stable set polytope. These results were discovered through a new connection that we make between one-way quantum communication protocols and semidefinite programming reformulations of LPs.

Reservoir Computing is a novel computing paradigm which uses a nonlinear recurrent dynamical system to carry out information processing. Here we will present an optoelectronic and an all-optical implementation of Reservoir Computers based on an architecture with a single nonlinear node and a delay loop. Moreover we will present an optical analogue readout which makes our reservoir computer a potential standalone solution after training. Our works show that, within the Reservoir Computing paradigm, all-optical computing with state-of-the-art performance is possible as a self-contained solution.

Reservoir computing is a new, powerful and flexible machine learning technique that is easily implemented in hardware. Recently, by using a time-multiplexed architecture, hardware reservoir computers have reached performance comparable to digital implementations. Operating speeds allowing for real time information operation have been reached using optoelectronic systems. At present the main performance bottleneck is the readout layer which uses slow, digital postprocessing. We have designed an analog readout suitable for time-multiplexed optoelectronic reservoir computers, capable of working in real time. The readout has been built and tested experimentally on a standard benchmark task. Its performance is better than non-reservoir methods, with ample room for further improvement. The present work thereby overcomes one of the major limitations for the future development of hardware reservoir computers.

We solve a 20-year old problem posed by Yannakakis and prove that there exists no polynomial-size linear program (LP) whose associated polytope projects to the traveling salesman polytope, even if the LP is not required to be symmetric. Moreover, we prove that this holds also for the cut polytope and the stable set polytope. These results were discovered through a new connection that we make between one-way quantum communication protocols and semidefinite programming reformulations of LPs.