Article
Chemistry, Multidisciplinary
Yutaro Yamazaki, Kentaro Kinoshita
Summary: This study develops a memristor that can respond to both electrical and optical stimuli, and demonstrates that the timescale of the transient current response of the device can be controlled by applying a small voltage. The computational performance of the device as a physical reservoir is evaluated in an image classification task, showing the potential for optimizing learning accuracy by tuning the device characteristics.
Article
Optics
Sendy Phang
Summary: Artificial intelligence (AI) drives the creation of disruptive future technologies that change the way we live and work. A new computing platform based on the photonic reservoir computing architecture, exploiting the dynamics of stimulated Brillouin scattering, is reported here. This passive optical system is suitable for use with high performance optical techniques, enabling real-time AI.
Article
Engineering, Mechanical
Yigong Yang, Pei Zhou, Penghua Mu, Nianqiang Li
Summary: This paper presents the first numerical implementation of photonic reservoir computing based on a spin VCSEL. The proposed system demonstrates fast response and has the potential to achieve high-speed information processing and lower power consumption.
NONLINEAR DYNAMICS
(2022)
Article
Chemistry, Physical
Nestor Ghenzi, Tae Won Park, Seung Soo Kim, Hae Jin Kim, Yoon Ho Jang, Kyung Seok Woo, Cheol Seong Hwang
Summary: This study experimentally and numerically investigates multiple switching modes in a Ta2O5/HfO2 memristor through reservoir computing (RC) simulation, revealing the significance of nonlinearity and heterogeneity in the RC framework. Unlike previous studies that employed homogeneous reservoirs, heterogeneity is introduced by combining different behaviors of the memristor units. The findings demonstrate the importance of these factors in improving pattern recognition performance in heterogeneous memristor RC systems with similar physical structures.
NANOSCALE HORIZONS
(2023)
Article
Computer Science, Artificial Intelligence
Piotr Antonik, Nicolas Marsal, Daniel Brunner, Damien Rontani
Summary: Reservoir computing is a growing paradigm for training recurrent neural networks. This work proposes the use of Bayesian optimization for efficient exploration of hyper-parameter space in a large-scale photonic system, resulting in notable improvements in performance.
COGNITIVE COMPUTATION
(2023)
Article
Biochemical Research Methods
Sebastian Freidank, Alfred Vogel, Norbert Linz
Summary: This study investigates the mechanisms of laser dissection in cornea using ultra-high-speed photography and finds that the cutting process relies on crack propagation along the bubble lobes. These insights are important for improving the cutting mechanisms in refractive surgery.
BIOMEDICAL OPTICS EXPRESS
(2022)
Article
Instruments & Instrumentation
Benjamin van Elburg, Gonzalo Collado-Lara, Gert-Wim Bruggert, Tim Segers, Michel Versluis, Guillaume Lajoinie
Summary: The study introduces a standalone lab-on-a-chip instrument that can efficiently produce monodisperse lipid-coated microbubbles, maintaining long-term stable, controlled, and safe operation using optical transmission-based measurement technique and feedback control method. This system demonstrates the ability to control bubble size and production rate, aiding in the development of microfluidic platforms for non-specialist end users.
REVIEW OF SCIENTIFIC INSTRUMENTS
(2021)
Article
Chemistry, Physical
Gianluca Milano, Giacomo Pedretti, Kevin Montano, Saverio Ricci, Shahin Hashemkhani, Luca Boarino, Daniele Ielmini, Carlo Ricciardi
Summary: By using a combination of self-organized nanowire networks and a memristive read-out layer, a hardware implementation of reservoir computing for recognition of spatio-temporal patterns and time-series prediction is demonstrated.
Article
Physics, Multidisciplinary
Jael Pauwels, Guy Van der Sande, Guy Verschaffelt, Serge Massar
Summary: The method proposed involves increasing the number of output neurons while keeping the reservoir fixed to improve the performance of a reservoir computer. It has been demonstrated to be effective in partially recovering lost performance in experimental opto-electronic systems subject to slow parameter drift. The scheme offers a trade-off between performance gains and system complexity.
Article
Nanoscience & Nanotechnology
Ian Bauwens, Krishan Harkhoe, Peter Bienstman, Guy Verschaffelt, Guy Van der Sande
Summary: This study proposes using transfer learning to address the issue of parameter drift in photonic reservoir computing system and reduce the resources required for retraining. Numerical studies on a delay-based system with semiconductor lasers demonstrate that transfer learning can mitigate parameter fluctuations and reduce training requirements for the second task.
Article
Nanoscience & Nanotechnology
Sarah Masaad, Emmanuel Gooskens, Stijn Sackesyn, Joni Dambre, Peter Bienstman
Summary: This paper simulates the use of a photonic reservoir to address nonlinearity-induced errors in fiber communication systems. By utilizing a KK receiver for direct detection and training the readout weights through backpropagation, successful compensation for nonlinear effects is achieved. Experimental results show that the photonic reservoir outperforms traditional optical feed-forward equalizer in terms of bit error rate, making it suitable for applications in data center communications.
Article
Mathematics, Interdisciplinary Applications
Lijun Pei, Mengyu Zhang
Summary: This paper explores the dynamics of a delay-based photonic reservoir computing system with a focus on double Hopf bifurcation. The existence of double Hopf bifurcation points is analyzed and bifurcation diagrams are drawn using DDE-BIFTOOL. Three types of double Hopf bifurcations are found, leading to stable equilibrium, stable periodic, and quasi-periodic solutions in distinct regions. These rich dynamical phenomena can aid in selecting suitable parameter values for optimal performance of the photonic reservoir computing system.
INTERNATIONAL JOURNAL OF BIFURCATION AND CHAOS
(2022)
Article
Optics
Yu Huang, Pei Zhou, Yigong Yang, Nianqiang Li
Summary: We proposed a high-speed photonic reservoir computing system using a compact Fano laser, which has a wider dynamic steady-state region and can achieve robust RC performance. The system has the potential applications for integrated neuromorphic photonic systems.
Article
Chemistry, Multidisciplinary
Alisina Bazrafshan, Maria-Eleni Kyriazi, Brandon Alexander Holt, Wenxiao Deng, Selma Piranej, Hanquan Su, Yuesong Hu, Afaf H. El-Sagheer, Tom Brown, Gabriel A. Kwong, Antonios G. Kanaras, Khalid Salaita
Summary: Research has shown that optimizing DNA motor performance can be achieved by adjusting structural parameters and buffer conditions. Increasing DNA leg density can improve speed and processivity, while DNA leg span can increase processivity and directionality. Label-free imaging has also revealed the unique motion patterns of the motors.
Article
Automation & Control Systems
Songlin Du, Ziwei Dong, Yuan Li, Takeshi Ikenaga
Summary: This article proposes a hardware-friendly Hough transform that can detect straight lines at ultrahigh speed. The key features include parallel processing of multiple pixels, direct calculation of line parameters, and simultaneous initialization and voting in the Hough parameter space without delay. Experimental results demonstrate real-time performance with a high frame rate of 784 frames/s and an ultralow delay of 0.7749 ms/frame.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Optics
Johnny Moughames, Xavier Porte, Michael Thiel, Gwenn Ulliac, Laurent Larger, Maxime Jacquot, Muamer Kadic, Daniel Brunner
Article
Automation & Control Systems
Antoine N. Andre, Patrick Sandoz, Benjamin Mauze, Maxime Jacquot, Guillaume J. Laurent
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2020)
Article
Energy & Fuels
Rania Mezzi, Nadia Yousfi-Steiner, Marie Cecile Pera, Daniel Hissel, Laurent Larger
Summary: This paper explores a fuel cell prognostics approach based on Echo State Network, which can be used for prediction under variable loads without prior knowledge, providing accurate prediction results.
Correction
Mathematics, Applied
Vladimir V. Semenov, Xavier Porte, Ibrahim Abdulhalim, Laurent Larger, Daniel Brunner
Article
Mathematics, Applied
Vladimir V. Semenov, Xavier Porte, Ibrahim Abdulhalim, Laurent Larger, Daniel Brunner
Summary: Nonlinear spatiotemporal systems are the basis for numerous physical phenomena in various fields, and the normal form description provides a canonical approach to unify models from different domains. Continuous transition between different types of dynamical systems by tuning accessible system parameters is highly relevant and can be achieved using an experimental platform.
Article
Computer Science, Artificial Intelligence
Nadezhda Semenova, Laurent Larger, Daniel Brunner
Summary: Deep neural networks have unlocked new applications previously reserved for higher human intelligence, leveraging computing power from special purpose hardware. However, the emulation of neural networks by binary computing leads to unsustainable energy consumption and slow speed. Research shows that noise accumulation in deep neural networks with noisy nonlinear neurons is generally limited, and noise can be completely suppressed when neuron activation functions have a slope smaller than unity.
Article
Mathematics, Interdisciplinary Applications
Daniel Brunner, Laurent Larger, Miguel C. Soriano
Summary: Driven by recent breakthroughs, photonics neural networks have seen a resurgence. This article provides an overview of progress in the past decade and outlines future developments. It focuses on photonic implementations of reservoir computing and discusses advances and challenges in implementing deep neural networks using photonics.
IEICE NONLINEAR THEORY AND ITS APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
A. N. Andre, P. Sandoz, M. Jacquot, G. J. Laurent
Summary: This paper discusses the problem of microscale pose estimation and presents a solution based on planar periodic targets. By utilizing Fourier spectrum analysis and long focal lengths, accurate estimation of the object's pose can be achieved with high resolution.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2022)
Article
Optics
Stephane Cuenat, Louis Andreoli, Antoine N. Andre, Patrick Sandoz, Guillaume J. Laurent, Raphael Couturier, Maxime Jacquot
Summary: The numerical wavefront backpropagation principle of digital holography offers extended focus capabilities without mechanical displacements. The authors propose a deep learning solution for autofocusing and test it on experimental and simulated holograms. The results show that the proposed tiny networks accurately infer the focusing distance with a short inference time.
Article
Optics
Pierre-Ambroise Lacourt, Francois Courvoisier, Jassem Safioui, Souleymane Diallo, Romain Martinenghi, Luca Furfaro, Maxime Jacquot, Jean-Marc Merolla, Luc Froehly, Laurent Larger
Summary: In this study, we present an improved manufacturing technique using femtosecond laser ablation to produce high-performance millimeter-sized whispering gallery mode resonators in calcium fluoride. The processing time is reduced by half compared to traditional methods. Our findings are supported by optical measurements and can be extended to other substrate materials.
Proceedings Paper
Engineering, Electrical & Electronic
Johnny Moughames, Xavier Porte, Laurent Larger, Maxime Jacquot, Muamer Kadic, Daniel Brunner
Summary: This paper presents scalable 3D photonic waveguide interconnects fabricated using two-photon polymerization, featuring optical waveguide couplers with 1.2 μm diameter and various branching topologies. A 225 input and 529 output interconnect is demonstrated as the final outcome.
2021 CONFERENCE ON LASERS AND ELECTRO-OPTICS (CLEO)
(2021)
Proceedings Paper
Engineering, Electrical & Electronic
Xavier Porte, Johnny Moughames, Laurent Larger, Maxime Jacquot, Muamer Kadic, Daniel Brunner
2021 CONFERENCE ON LASERS AND ELECTRO-OPTICS EUROPE & EUROPEAN QUANTUM ELECTRONICS CONFERENCE (CLEO/EUROPE-EQEC)
(2021)
Article
Engineering, Electrical & Electronic
Antoine N. Andre, Patrick Sandoz, Benjamin Mauze, Maxime Jacquot, Guillaume J. Laurent
Summary: This study examines the robustness of a flat position measurement method based on a pseudo-periodic pattern, which can be implemented at different scales with unique capabilities that include high resolution, large measurement range, and resistance to various types of disturbances.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)