On-chip diffractive optical neural networks(DONNs)bring the advantages of parallel processing and low energy consumption.However,an accurate representation of the optical field’s evolution in the structure cannot be ...On-chip diffractive optical neural networks(DONNs)bring the advantages of parallel processing and low energy consumption.However,an accurate representation of the optical field’s evolution in the structure cannot be provided using the previous diffraction-based analysis method.Moreover,the loss caused by the open boundaries poses challenges to applications.A multimode DONN architecture based on a more precise eigenmode analysis method is proposed.We have constructed a universal library of input,output,and metaline structures utilizing this method,and realized a multimode DONN composed of the structures from the library.On the designed multimode DONNs with only one layer of the metaline,the classification task of an Iris plants dataset is verified with an accuracy of 90%on the blind test dataset,and the performance of the one-bit binary adder task is also validated.Compared to the previous architectures,the multimode DONN exhibits a more compact design and higher energy efficiency.展开更多
Parallel multi-thread processing in advanced intelligent processors is the core to realize high-speed and high-capacity signal processing systems.Optical neural network(ONN)has the native advantages of high paralleliz...Parallel multi-thread processing in advanced intelligent processors is the core to realize high-speed and high-capacity signal processing systems.Optical neural network(ONN)has the native advantages of high parallelization,large bandwidth,and low power consumption to meet the demand of big data.Here,we demonstrate the dual-layer ONN with Mach-Zehnder interferometer(MZI)network and nonlinear layer,while the nonlinear activation function is achieved by optical-electronic signal conversion.Two frequency components from the microcomb source carrying digit datasets are simultaneously imposed and intelligently recognized through the ONN.We successfully achieve the digit classification of different frequency components by demultiplexing the output signal and testing power distribution.Efficient parallelization feasibility with wavelength division multiplexing is demonstrated in our high-dimensional ONN.This work provides a high-performance architecture for future parallel high-capacity optical analog computing.展开更多
Optical neural networks(ONNs)are a class of emerging computing platforms that leverage the properties of light to perform ultra-fast computations with ultra-low energy consumption.ONNs often use CCD cameras as the out...Optical neural networks(ONNs)are a class of emerging computing platforms that leverage the properties of light to perform ultra-fast computations with ultra-low energy consumption.ONNs often use CCD cameras as the output layer.In this work,we propose the use of perovskite solar cells as a promising alternative to imaging cameras in ONN designs.Solar cells are ubiquitous,versatile,highly customizable,and can be fabricated quickly in laboratories.Their large acquisition area and outstanding efficiency enable them to generate output signals with a large dynamic range without the need for amplification.Here we have experimentally demonstrated the feasibility of using perovskite solar cells for capturing ONN output states,as well as the capability of single-layer random ONNs to achieve excellent performance even with a very limited number of pixels.Our results show that the solar-cell-based ONN setup consistently outperforms the same setup with CCD cameras of the same resolution.These findings highlight the potential of solar-cell-based ONNs as an ideal choice for automated and battery-free edge-computing applications.展开更多
Deep learning has rapidly advanced amidst the proliferation of large models,leading to challenges in computational resources and power consumption.Optical neural networks(ONNs)offer a solution by shifting computation ...Deep learning has rapidly advanced amidst the proliferation of large models,leading to challenges in computational resources and power consumption.Optical neural networks(ONNs)offer a solution by shifting computation to optics,thereby leveraging the benefits of low power consumption,low latency,and high parallelism.The current training paradigm for ONNs primarily relies on backpropagation(BP).However,the reliance is incompatible with potential unknown processes within the system,which necessitates detailed knowledge and precise mathematical modeling of the optical process.In this paper,we present a pre-sensor multilayer ONN with nonlinear activation,utilizing a forward-forward algorithm to directly train both optical and digital parameters,which replaces the traditional backward pass with an additional forward pass.Our proposed nonlinear optical system demonstrates significant improvements in image classification accuracy,achieving a maximum enhancement of 9.0%.It also validates the efficacy of training parameters in the presence of unknown nonlinear components in the optical system.The proposed training method addresses the limitations of BP,paving the way for applications with a broader range of physical transformations in ONNs.展开更多
Mode-division multiplexing technology has been proposed as a crucial technique for enhancing communication capacity and alleviating growing communication demands.Optical switching,which is an essential component of op...Mode-division multiplexing technology has been proposed as a crucial technique for enhancing communication capacity and alleviating growing communication demands.Optical switching,which is an essential component of optical communication systems,enables information exchange between channels.However,existing optical switching solutions are inadequate for addressing flexible information exchange among the mode channels.In this study,we introduced a flexible mode switching system in a multimode fibre based on an optical neural network chip.This system utilised the flexibility of on-chip optical neural networks along with an all-fibre orbital angular momentum(OAM)mode multiplexer-demultiplexer to achieve mode switching among the three OAM modes within a multimode fibre.The system adopted an improved gradient descent algorithm to achieve training for arbitrary 3×3 exchange matrices and ensured maximum crosstalk of less than-18.7 dB,thus enabling arbitrary inter-mode channel information exchange.The proposed optical-neural-network-based mode-switching system was experimentally validated by successfully transmitting different modulation formats across various modes.This innovative solution holds promise for providing effective optical switching in practical multimode communication networks.展开更多
Artificial neural networks have dramatically improved the performance of many machine-learning applications such as image recognition and natural language processing. However, the electronic hardware implementations o...Artificial neural networks have dramatically improved the performance of many machine-learning applications such as image recognition and natural language processing. However, the electronic hardware implementations of the above-mentioned tasks are facing performance ceiling because Moore’s Law is slowing down. In this article, we propose an optical neural network architecture based on optical scattering units to implement deep learning tasks with fast speed, low power consumption and small footprint.The optical scattering units allow light to scatter back and forward within a small region and can be optimized through an inverse design method. The optical scattering units can implement high-precision stochastic matrix multiplication with mean squared error < 10-4 and a mere 4*4 um2 footprint.Furthermore, an optical neural network framework based on optical scattering units is constructed by introducing "Kernel Matrix", which can achieve 97.1% accuracy on the classic image classification dataset MNIST.展开更多
Optical neural networks (ONNs), enabling low latency and high parallel data processing withoutelectromagnetic interference, have become a viable player for fast and energy-efficient processing andcalculation to meet t...Optical neural networks (ONNs), enabling low latency and high parallel data processing withoutelectromagnetic interference, have become a viable player for fast and energy-efficient processing andcalculation to meet the increasing demand for hash rate. Photonic memories employing nonvolatile phase-change materials could achieve zero static power consumption, low thermal cross talk, large-scale, andhigh-energy-efficient photonic neural networks. Nevertheless, the switching speed and dynamic energyconsumption of phase-change material-based photonic memories make them inapplicable for in situ training.Here, by integrating a patch of phase change thin film with a PIN-diode-embedded microring resonator,a bifunctional photonic memory enabling both 5-bit storage and nanoseconds volatile modulation wasdemonstrated. For the first time, a concept is presented for electrically programmable phase-changematerial-driven photonic memory integrated with nanosecond modulation to allow fast in situ training and zerostatic power consumption data processing in ONNs. ONNs with an optical convolution kernel constructedby our photonic memory theoretically achieved an accuracy of predictions higher than 95% when testedby the MNIST handwritten digit database. This provides a feasible solution to constructing large-scalenonvolatile ONNs with high-speed in situ training capability.展开更多
With the advent of the era of big data,artificial intelligence has attracted continuous attention from all walks of life,and has been widely used in medical image analysis,molecular and material science,language recog...With the advent of the era of big data,artificial intelligence has attracted continuous attention from all walks of life,and has been widely used in medical image analysis,molecular and material science,language recognition and other fields.As the basis of artificial intelligence,the research results of neural network are remarkable.However,due to the inherent defect that electrical signal is easily interfered and the processing speed is proportional to the energy loss,researchers have turned their attention to light,trying to build neural networks in the field of optics,making full use of the parallel processing ability of light to solve the problems of electronic neural networks.After continuous research and development,optical neural network has become the forefront of the world.Here,we mainly introduce the development of this field,summarize and compare some classical researches and algorithm theories,and look forward to the future of optical neural network.展开更多
Optical computing provides unique opportunities in terms of parallelization,scalability,power efficiency,and computational speed and has attracted major interest for machine learning.Diffractive deep neural networks h...Optical computing provides unique opportunities in terms of parallelization,scalability,power efficiency,and computational speed and has attracted major interest for machine learning.Diffractive deep neural networks have been introduced earlier as an optical machine learning framework that uses task-specific diffractive surfaces designed by deep learning to all-optically perform inference,achieving promising performance for object classification and imaging.We demonstrate systematic improvements in diffractive optical neural networks,based on a differential measurement technique that mitigates the strict nonnegativity constraint of light intensity.In this differential detection scheme,each class is assigned to a separate pair of detectors,behind a diffractive optical network,and the class inference is made by maximizing the normalized signal difference between the photodetector pairs.Using this differential detection scheme,involving 10 photodetector pairs behind 5 diffractive layers with a total of 0.2 million neurons,we numerically achieved blind testing accuracies of 98.54%,90.54%,and 48.51%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively.Moreover,by utilizing the inherent parallelization capability of optical systems,we reduced the cross-talk and optical signal coupling between the positive and negative detectors of each class by dividing the optical path into two jointly trained diffractive neural networks that work in parallel.We further made use of this parallelization approach and divided individual classes in a target dataset among multiple jointly trained diffractive neural networks.Using this class-specific differential detection in jointly optimized diffractive neural networks that operate in parallel,our simulations achieved blind testing accuracies of 98.52%,91.48%,and 50.82%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively,coming close to the performance of some of the earlier generations of all-electronic deep neural networks,e.g.,LeNet,which achieves classification accuracies of 98.77%,90.27%,and 55.21%corresponding to the same datasets,respectively.In addition to these jointly optimized diffractive neural networks,we also independently optimized multiple diffractive networks and utilized them in a way that is similar to ensemble methods practiced in machine learning;using 3 independently optimized differential diffractive neural networks that optically project their light onto a common output/detector plane,we numerically achieved blind testing accuracies of 98.59%,91.06%,and 51.44%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively.Through these systematic advances in designing diffractive neural networks,the reported classification accuracies set the state of the art for all-optical neural network design.The presented framework might be useful to bring optical neural network-based low power solutions for various machine learning applications and help us design new computational cameras that are task-specific.展开更多
Artificial neural network with broad application prospect has attracted particular attention due to the promise of solving the memory wall bottleneck.The neural devices that mix light and electricity provide more degr...Artificial neural network with broad application prospect has attracted particular attention due to the promise of solving the memory wall bottleneck.The neural devices that mix light and electricity provide more degrees of freedom for the design of artificial neural network,but they still do not get rid of the shackles that the response signal needs circuit to transmission.The exploration of all-optical neural devices(optical signal input and output)is expected to solve this problem.Here,an all-optical synaptic device simply based on a long-afterglow material is reported.The optical properties of the all-optical synaptic device are similar to the responses in biological synapses.Unique image displays and memory functions can be achieved by combining alloptical synaptic arrays with synaptic memory behavior.Furthermore,the optical summation of all-optical synaptic array pixels can be completed by combining the focusing characteristics of convex lens,which realizes the photon transmission after preprocessing multiple input signals.Particularly,the simple single-layer structure of all-optical synapses with polydimethylsiloxane(PDMS)as the carrier has high plasticity and is expected to achieve large-scale preparation.This work enriches the diversity of artificial synapses and shows the huge development potential of photoelectric artificial neural networks.展开更多
Integrated diffractive optical neural networks(DONNs)have significant potential for complex machine learning tasks with high speed and ultralow energy consumption.However,the on-chip implementation of a high-performan...Integrated diffractive optical neural networks(DONNs)have significant potential for complex machine learning tasks with high speed and ultralow energy consumption.However,the on-chip implementation of a high-performance optical neural network is limited by input dimensions.In contrast to existing photonic neural networks,a space-time interleaving technology based on arrayed waveguides is designed to realize an on-chip DONN with high-speed,high-dimensional,and all-optical input signal modulation.To demonstrate the performance of the on-chip DONN with high-speed space-time interleaving modulation,an on-chip DONN with a designed footprint of 0.0945 mm~2is proposed to resolve the vowel recognition task,reaching a computation speed of about 1.4×10^(13)operations per second and yielding an accuracy of 98.3%in numerical calculation.In addition,the function of the specially designed arrayed waveguides for realizing parallel signal inputs using space-time conversion has been verified experimentally.This method can realize the on-chip DONN with higher input dimension and lower energy consumption.展开更多
Quantum state tomography(QST)is a crucial ingredient for almost all aspects of experimental quantum information processing.As an analog of the“imaging”technique in quantum settings,QST is born to be a data science p...Quantum state tomography(QST)is a crucial ingredient for almost all aspects of experimental quantum information processing.As an analog of the“imaging”technique in quantum settings,QST is born to be a data science problem,where machine learning techniques,noticeably neural networks,have been applied extensively.We build and demonstrate an optical neural network(ONN)for photonic polarization qubit QST.The ONN is equipped with built-in optical nonlinear activation functions based on electromagnetically induced transparency.The experimental results show that our ONN can determine the phase parameter of the qubit state accurately.As optics are highly desired for quantum interconnections,our ONN-QST may contribute to the realization of optical quantum networks and inspire the ideas combining artificial optical intelligence with quantum information studies.展开更多
Optical neural network(ONNs)are emerging as attractive propos-als for machine-learning applications.However,the stability of ONNs decreases with the circuit depth,limiting the scalability of ONNs for practical uses.He...Optical neural network(ONNs)are emerging as attractive propos-als for machine-learning applications.However,the stability of ONNs decreases with the circuit depth,limiting the scalability of ONNs for practical uses.Here we demonstrate how to compress the circuit depth to scale only logarithmically in terms of the dimension of the data,leading to an exponential gain in terms of noise robustness.Our low-depth(LD)-ONN is based on an architecture,called Optical Com-puTing Of dot-Product UnitS(OCTOPUS),which can also be applied individually as a linear perceptron for solving classification problems.We present both numerical and theoretical evidence showing that LD-ONN can exhibit a significant improvement on robustness,compared with previous ONN proposals based on singular-value decomposition.展开更多
Optics is an exciting route for the next generation of computing hardware for machine learning,promising several orders of magnitude enhancement in both computational speed and energy efficiency.However, reaching the ...Optics is an exciting route for the next generation of computing hardware for machine learning,promising several orders of magnitude enhancement in both computational speed and energy efficiency.However, reaching the full capacity of an optical neural network(NN) necessitates that the computing be implemented optically not only for inference but also for training. The primary algorithm for network training is backpropagation, in which the calculation is performed in the order opposite to the information flow for inference. Although straightforward in a digital computer, the optical implementation of backpropagation has remained elusive, particularly because of the conflicting requirements for the optical element that implements the nonlinear activation function. We address this challenge for the first time, we believe, with a surprisingly simple scheme, employing saturable absorbers for the role of activation units. Our approach is adaptable to various analog platforms and materials and demonstrates the possibility of constructing NNs entirely reliant on analog optical processes for both training and inference tasks.展开更多
Optical deep learning based on diffractive optical elements offers unique advantages for parallel processing,computational speed,and power efficiency.One landmark method is the diffractive deep neural network(D^(2) NN...Optical deep learning based on diffractive optical elements offers unique advantages for parallel processing,computational speed,and power efficiency.One landmark method is the diffractive deep neural network(D^(2) NN)based on three-dimensional printing technology operated in the terahertz spectral range.Since the terahertz bandwidth involves limited interparticle coupling and material losses,this paper extends D^(2) NN to visible wavelengths.A general theory including a revised formula is proposed to solve any contradictions between wavelength,neuron size,and fabrication limitations.A novel visible light D^(2) NN classifier is used to recognize unchanged targets(handwritten digits ranging from 0 to 9)and targets that have been changed(i.e.,targets that have been covered or altered)at a visible wavelength of 632.8 nm.The obtained experimental classification accuracy(84%)and numerical classification accuracy(91.57%)quantify the match between the theoretical design and fabricated system performance.The presented framework can be used to apply a D^(2) NN to various practical applications and design other new applications.展开更多
The explosive growth of data and information has motivated various emerging non-von Neumann computational approaches in the More-than-Moore era.Photonics neuromorphic computing has attracted lots of attention due to t...The explosive growth of data and information has motivated various emerging non-von Neumann computational approaches in the More-than-Moore era.Photonics neuromorphic computing has attracted lots of attention due to the fascinating advantages such as high speed,wide bandwidth,and massive parallelism.Here,we offer a review on the optical neural computing in our research groups at the device and system levels.The photonics neuron and photonics synapse plasticity are presented.In addition,we introduce several optical neural computing architectures and algorithms including photonic spiking neural network,photonic convolutional neural network,photonic matrix computation,photonic reservoir computing,and photonic reinforcement learning.Finally,we summarize the major challenges faced by photonic neuromorphic computing,and propose promising solutions and perspectives.展开更多
Artificial intelligence(AI)has taken breathtaking leaps forward in recent years,evolving into a strategic technology for pioneering the future.The growing demand for computing power—especially in demanding inference ...Artificial intelligence(AI)has taken breathtaking leaps forward in recent years,evolving into a strategic technology for pioneering the future.The growing demand for computing power—especially in demanding inference tasks,exemplified by generative AI models such as ChatGPT—poses challenges for conventional electronic computing systems.Advances in photonics technology have ignited interest in investigating photonic computing as a promising AI computing modality.Through the profound fusion of AI and photonics technologies,intelligent photonics is developing as an emerging interdisciplinary field with significant potential to revolutionize practical applications.Deep learning,as a subset of AI,presents efficient avenues for optimizing photonic design,developing intelligent optical systems,and performing optical data processing and analysis.Employing AI in photonics can empower applications such as smartphone cameras,biomedical microscopy,and virtual and augmented reality displays.Conversely,leveraging photonics-based devices and systems for the physical implementation of neural networks enables high speed and low energy consumption.Applying photonics technology in AI computing is expected to have a transformative impact on diverse fields,including optical communications,automatic driving,and astronomical observation.Here,recent advances in intelligent photonics are presented from the perspective of the synergy between deep learning and metaphotonics,holography,and quantum photonics.This review also spotlights relevant applications and offers insights into challenges and prospects.展开更多
Optical neural networks(ONNs)offer a promising solution for high-performance,energy-efficient artificial intelligence hardware by leveraging the parallelism and speed of light.However,the large-scale implementation of...Optical neural networks(ONNs)offer a promising solution for high-performance,energy-efficient artificial intelligence hardware by leveraging the parallelism and speed of light.However,the large-scale implementation of ONNs remains challenging due to the bulky footprint and complex control of optical synapses.In this work,we propose and simulate a plasmonic polarized synaptic architecture that overcomes the diffraction limit and enables ultra-compact ONNs.By tuning the polarization state of incident light,the optical transmittance through each plasmonic unit can be dynamically adjusted to represent a synaptic weight.Our plasmonic structures,with features as small as 40 nm,operate well below this limit in the visible spectrum(400-750 nm).Compared with diffraction and interference-based circuit designs,our proposed method achieves a substantial reduction in synaptic density by factors of 150000-fold and 1500-fold,respectively.Furthermore,we successfully demonstrate a proof-of-concept plasmonic ONN applied to the Canadian Institute for Advanced Research—10 classes(CIFAR-10)dataset using a Visual Geometry Group network with 16 layers(VGG16)model.After training for 80 epochs,the network achieves an accuracy of 93%.The polarization-tunable plasmonics paves the way towards scalable ONNs for next-generation artificial intelligence(AI)accelerators and smart sensors.展开更多
Optical neural networks have significant advantages in terms of power consumption,parallelism,and high computing speed,which has intrigued extensive attention in both academic and engineering communities.It has been c...Optical neural networks have significant advantages in terms of power consumption,parallelism,and high computing speed,which has intrigued extensive attention in both academic and engineering communities.It has been considered as one of the powerful tools in promoting the fields of imaging processing and object recognition.However,the existing optical system architecture cannot be reconstructed to the realization of multi-functional artificial intelligence systems simultaneously.To push the development of this issue,we propose the pluggable diffractive neural networks(P-DNN),a general paradigm resorting to the cascaded metasurfaces,which can be applied to recognize various tasks by switching internal plug-ins.As the proof-of-principle,the recognition functions of six types of handwritten digits and six types of fashions are numerical simulated and experimental demonstrated at near-infrared regimes.Encouragingly,the proposed paradigm not only improves the flexibility of the optical neural networks but paves the new route for achieving high-speed,low-power and versatile artificial intelligence systems.展开更多
The rapid growth of deep learning applications has sparked a revolution in computing paradigms,with optical neural networks(ONNs)emerging as a promising platform for achieving ultra-high computing power and energy eff...The rapid growth of deep learning applications has sparked a revolution in computing paradigms,with optical neural networks(ONNs)emerging as a promising platform for achieving ultra-high computing power and energy efficiency.Despite great progress in analog optical computing,the lack of scalable optical nonlinearities and losses in photonic devices pose considerable challenges for power levels,energy efficiency,and signal latency.Here,we report an end-to-end all-optical nonlinear activator that utilizes the energy conversion of Brillouin scattering to perform efficient nonlinear processing.The activator exhibits an ultra-low activation threshold(24 nW),a wide transmission bandwidth(over 40 GHz),strong robustness,and high energy transfer efficiency.These advantages provide a feasible solution to overcome the existing bottlenecks in ONNs.As a proof-of-concept,a series of tasks is designed to validate the capability of the proposed activator as an activation unit for ONNs.Simulations show that the experiment-based nonlinear model outperforms classical activation functions in classification(97.64%accuracy for MNIST and 87.84%for Fashion-MNIST)and regression(with a symbol error rate as low as 0%)tasks.This work provides valuable insights into the innovative design of all-optical neural networks.展开更多
基金supported by the National Natural Science Foundation of China (Grant No.62135009)the Beijing Municipal Science and Technology Commission,Administrative Commission of Zhongguancun Science Park (Grant No.Z221100005322010).
文摘On-chip diffractive optical neural networks(DONNs)bring the advantages of parallel processing and low energy consumption.However,an accurate representation of the optical field’s evolution in the structure cannot be provided using the previous diffraction-based analysis method.Moreover,the loss caused by the open boundaries poses challenges to applications.A multimode DONN architecture based on a more precise eigenmode analysis method is proposed.We have constructed a universal library of input,output,and metaline structures utilizing this method,and realized a multimode DONN composed of the structures from the library.On the designed multimode DONNs with only one layer of the metaline,the classification task of an Iris plants dataset is verified with an accuracy of 90%on the blind test dataset,and the performance of the one-bit binary adder task is also validated.Compared to the previous architectures,the multimode DONN exhibits a more compact design and higher energy efficiency.
基金Peng Xie acknowledges the support from the China Scholarship Council(Grant no.201804910829).
文摘Parallel multi-thread processing in advanced intelligent processors is the core to realize high-speed and high-capacity signal processing systems.Optical neural network(ONN)has the native advantages of high parallelization,large bandwidth,and low power consumption to meet the demand of big data.Here,we demonstrate the dual-layer ONN with Mach-Zehnder interferometer(MZI)network and nonlinear layer,while the nonlinear activation function is achieved by optical-electronic signal conversion.Two frequency components from the microcomb source carrying digit datasets are simultaneously imposed and intelligently recognized through the ONN.We successfully achieve the digit classification of different frequency components by demultiplexing the output signal and testing power distribution.Efficient parallelization feasibility with wavelength division multiplexing is demonstrated in our high-dimensional ONN.This work provides a high-performance architecture for future parallel high-capacity optical analog computing.
基金support from the Carnegie Trust for the Universities of ScotlandPRIN 2022597MBS PHERMIACsupported by the European Research Council(ERC)under the European Union Horizon 2020 Research and Innovation Program(Grant Agreement No.819346)。
文摘Optical neural networks(ONNs)are a class of emerging computing platforms that leverage the properties of light to perform ultra-fast computations with ultra-low energy consumption.ONNs often use CCD cameras as the output layer.In this work,we propose the use of perovskite solar cells as a promising alternative to imaging cameras in ONN designs.Solar cells are ubiquitous,versatile,highly customizable,and can be fabricated quickly in laboratories.Their large acquisition area and outstanding efficiency enable them to generate output signals with a large dynamic range without the need for amplification.Here we have experimentally demonstrated the feasibility of using perovskite solar cells for capturing ONN output states,as well as the capability of single-layer random ONNs to achieve excellent performance even with a very limited number of pixels.Our results show that the solar-cell-based ONN setup consistently outperforms the same setup with CCD cameras of the same resolution.These findings highlight the potential of solar-cell-based ONNs as an ideal choice for automated and battery-free edge-computing applications.
基金National Key Research and Development Program of China(2024YFE0203600)National Natural Science Foundation of China(62135009)Beijing Municipal Science&Technology Commission,Administrative Commission of Zhongguancun Science Park(Z221100005322010)。
文摘Deep learning has rapidly advanced amidst the proliferation of large models,leading to challenges in computational resources and power consumption.Optical neural networks(ONNs)offer a solution by shifting computation to optics,thereby leveraging the benefits of low power consumption,low latency,and high parallelism.The current training paradigm for ONNs primarily relies on backpropagation(BP).However,the reliance is incompatible with potential unknown processes within the system,which necessitates detailed knowledge and precise mathematical modeling of the optical process.In this paper,we present a pre-sensor multilayer ONN with nonlinear activation,utilizing a forward-forward algorithm to directly train both optical and digital parameters,which replaces the traditional backward pass with an additional forward pass.Our proposed nonlinear optical system demonstrates significant improvements in image classification accuracy,achieving a maximum enhancement of 9.0%.It also validates the efficacy of training parameters in the presence of unknown nonlinear components in the optical system.The proposed training method addresses the limitations of BP,paving the way for applications with a broader range of physical transformations in ONNs.
基金supported by the National Natural Science Foundation of China(NSFC)(62125503,62261160388)Natural Science Foundation of Hubei Province of China(2023AFA028)+1 种基金Key R&D Program of Hubei Province of China(2020BAB001,2021BAA024)Innovation Project of Optics Valley Laboratory(OVL2021BG004).
文摘Mode-division multiplexing technology has been proposed as a crucial technique for enhancing communication capacity and alleviating growing communication demands.Optical switching,which is an essential component of optical communication systems,enables information exchange between channels.However,existing optical switching solutions are inadequate for addressing flexible information exchange among the mode channels.In this study,we introduced a flexible mode switching system in a multimode fibre based on an optical neural network chip.This system utilised the flexibility of on-chip optical neural networks along with an all-fibre orbital angular momentum(OAM)mode multiplexer-demultiplexer to achieve mode switching among the three OAM modes within a multimode fibre.The system adopted an improved gradient descent algorithm to achieve training for arbitrary 3×3 exchange matrices and ensured maximum crosstalk of less than-18.7 dB,thus enabling arbitrary inter-mode channel information exchange.The proposed optical-neural-network-based mode-switching system was experimentally validated by successfully transmitting different modulation formats across various modes.This innovative solution holds promise for providing effective optical switching in practical multimode communication networks.
基金This work was supported by the National Key Research and Development Program of China(2017YFA0205700)the National Natural Science Foundation of China(61927820)Yurui Qu was supported by Zhejiang Lab’s International Talent Fund for Young Professionals.
文摘Artificial neural networks have dramatically improved the performance of many machine-learning applications such as image recognition and natural language processing. However, the electronic hardware implementations of the above-mentioned tasks are facing performance ceiling because Moore’s Law is slowing down. In this article, we propose an optical neural network architecture based on optical scattering units to implement deep learning tasks with fast speed, low power consumption and small footprint.The optical scattering units allow light to scatter back and forward within a small region and can be optimized through an inverse design method. The optical scattering units can implement high-precision stochastic matrix multiplication with mean squared error < 10-4 and a mere 4*4 um2 footprint.Furthermore, an optical neural network framework based on optical scattering units is constructed by introducing "Kernel Matrix", which can achieve 97.1% accuracy on the classic image classification dataset MNIST.
基金supported by the National Key Research and Development Program of China (2019YFB2203002 and 2021YFB2801300)National Natural Science Foundation of China (62105287, 91950204, and 61975179)Zhejiang Provincial Natural Science Foundation (LD22F040002)
文摘Optical neural networks (ONNs), enabling low latency and high parallel data processing withoutelectromagnetic interference, have become a viable player for fast and energy-efficient processing andcalculation to meet the increasing demand for hash rate. Photonic memories employing nonvolatile phase-change materials could achieve zero static power consumption, low thermal cross talk, large-scale, andhigh-energy-efficient photonic neural networks. Nevertheless, the switching speed and dynamic energyconsumption of phase-change material-based photonic memories make them inapplicable for in situ training.Here, by integrating a patch of phase change thin film with a PIN-diode-embedded microring resonator,a bifunctional photonic memory enabling both 5-bit storage and nanoseconds volatile modulation wasdemonstrated. For the first time, a concept is presented for electrically programmable phase-changematerial-driven photonic memory integrated with nanosecond modulation to allow fast in situ training and zerostatic power consumption data processing in ONNs. ONNs with an optical convolution kernel constructedby our photonic memory theoretically achieved an accuracy of predictions higher than 95% when testedby the MNIST handwritten digit database. This provides a feasible solution to constructing large-scalenonvolatile ONNs with high-speed in situ training capability.
基金supported in part by the National Natural Science Foundation of China under Grant 11773018 and Grant 61727802in part by the Key Research and Development programs in Jiangsu China under Grant BE2018126+1 种基金in part by the Fundamental Research Funds for the Central Universities under Grant 30919011401 and Grant 30920010001in part by the Leading Technology of Jiangsu Basic Research Plan under Grant BK20192003.
文摘With the advent of the era of big data,artificial intelligence has attracted continuous attention from all walks of life,and has been widely used in medical image analysis,molecular and material science,language recognition and other fields.As the basis of artificial intelligence,the research results of neural network are remarkable.However,due to the inherent defect that electrical signal is easily interfered and the processing speed is proportional to the energy loss,researchers have turned their attention to light,trying to build neural networks in the field of optics,making full use of the parallel processing ability of light to solve the problems of electronic neural networks.After continuous research and development,optical neural network has become the forefront of the world.Here,we mainly introduce the development of this field,summarize and compare some classical researches and algorithm theories,and look forward to the future of optical neural network.
文摘Optical computing provides unique opportunities in terms of parallelization,scalability,power efficiency,and computational speed and has attracted major interest for machine learning.Diffractive deep neural networks have been introduced earlier as an optical machine learning framework that uses task-specific diffractive surfaces designed by deep learning to all-optically perform inference,achieving promising performance for object classification and imaging.We demonstrate systematic improvements in diffractive optical neural networks,based on a differential measurement technique that mitigates the strict nonnegativity constraint of light intensity.In this differential detection scheme,each class is assigned to a separate pair of detectors,behind a diffractive optical network,and the class inference is made by maximizing the normalized signal difference between the photodetector pairs.Using this differential detection scheme,involving 10 photodetector pairs behind 5 diffractive layers with a total of 0.2 million neurons,we numerically achieved blind testing accuracies of 98.54%,90.54%,and 48.51%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively.Moreover,by utilizing the inherent parallelization capability of optical systems,we reduced the cross-talk and optical signal coupling between the positive and negative detectors of each class by dividing the optical path into two jointly trained diffractive neural networks that work in parallel.We further made use of this parallelization approach and divided individual classes in a target dataset among multiple jointly trained diffractive neural networks.Using this class-specific differential detection in jointly optimized diffractive neural networks that operate in parallel,our simulations achieved blind testing accuracies of 98.52%,91.48%,and 50.82%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively,coming close to the performance of some of the earlier generations of all-electronic deep neural networks,e.g.,LeNet,which achieves classification accuracies of 98.77%,90.27%,and 55.21%corresponding to the same datasets,respectively.In addition to these jointly optimized diffractive neural networks,we also independently optimized multiple diffractive networks and utilized them in a way that is similar to ensemble methods practiced in machine learning;using 3 independently optimized differential diffractive neural networks that optically project their light onto a common output/detector plane,we numerically achieved blind testing accuracies of 98.59%,91.06%,and 51.44%for MNIST,Fashion-MNIST,and grayscale CIFAR-10 datasets,respectively.Through these systematic advances in designing diffractive neural networks,the reported classification accuracies set the state of the art for all-optical neural network design.The presented framework might be useful to bring optical neural network-based low power solutions for various machine learning applications and help us design new computational cameras that are task-specific.
基金The authors are grateful for financial support from the National Natural Science Foundation of China(No.U21A20497)the Natural Science Foundation for Distinguished Young Scholars of Fujian Province(No.2020J06012)+1 种基金the Fujian Science&Technology Innovation Laboratory for Optoelectronic Information of China(No.2021ZZ129)the Joint Funds for the innovation of science and Technology,Fujian province(No.2021Y9074).
文摘Artificial neural network with broad application prospect has attracted particular attention due to the promise of solving the memory wall bottleneck.The neural devices that mix light and electricity provide more degrees of freedom for the design of artificial neural network,but they still do not get rid of the shackles that the response signal needs circuit to transmission.The exploration of all-optical neural devices(optical signal input and output)is expected to solve this problem.Here,an all-optical synaptic device simply based on a long-afterglow material is reported.The optical properties of the all-optical synaptic device are similar to the responses in biological synapses.Unique image displays and memory functions can be achieved by combining alloptical synaptic arrays with synaptic memory behavior.Furthermore,the optical summation of all-optical synaptic array pixels can be completed by combining the focusing characteristics of convex lens,which realizes the photon transmission after preprocessing multiple input signals.Particularly,the simple single-layer structure of all-optical synapses with polydimethylsiloxane(PDMS)as the carrier has high plasticity and is expected to achieve large-scale preparation.This work enriches the diversity of artificial synapses and shows the huge development potential of photoelectric artificial neural networks.
基金supported by the National Natural Science Foundation of China(NSFC)(No.62135009)the Beijing Municipal Science&Technology Commission,Administrative Commission of Zhongguancun Science Park(No.Z221100005322010)。
文摘Integrated diffractive optical neural networks(DONNs)have significant potential for complex machine learning tasks with high speed and ultralow energy consumption.However,the on-chip implementation of a high-performance optical neural network is limited by input dimensions.In contrast to existing photonic neural networks,a space-time interleaving technology based on arrayed waveguides is designed to realize an on-chip DONN with high-speed,high-dimensional,and all-optical input signal modulation.To demonstrate the performance of the on-chip DONN with high-speed space-time interleaving modulation,an on-chip DONN with a designed footprint of 0.0945 mm~2is proposed to resolve the vowel recognition task,reaching a computation speed of about 1.4×10^(13)operations per second and yielding an accuracy of 98.3%in numerical calculation.In addition,the function of the specially designed arrayed waveguides for realizing parallel signal inputs using space-time conversion has been verified experimentally.This method can realize the on-chip DONN with higher input dimension and lower energy consumption.
基金supported by General Research Fund (No. GRF/16300220)
文摘Quantum state tomography(QST)is a crucial ingredient for almost all aspects of experimental quantum information processing.As an analog of the“imaging”technique in quantum settings,QST is born to be a data science problem,where machine learning techniques,noticeably neural networks,have been applied extensively.We build and demonstrate an optical neural network(ONN)for photonic polarization qubit QST.The ONN is equipped with built-in optical nonlinear activation functions based on electromagnetically induced transparency.The experimental results show that our ONN can determine the phase parameter of the qubit state accurately.As optics are highly desired for quantum interconnections,our ONN-QST may contribute to the realization of optical quantum networks and inspire the ideas combining artificial optical intelligence with quantum information studies.
基金supported by the Natural Science Foundation of Guangdong Province(Grant No.2017B030308003)the Key R&D Pro-gram of Guangdong province(Grant No.2018B030326001)+2 种基金the Sci-ence,Technology and Innovation Commission of Shenzhen Municipality(Grant No.JCYJ20170412152620376 and No.JCYJ20170817105046702 and No.KYTDPT20181011104202253)National Natural Science Foundation of China(Grant No.11875160 and No.U1801661)the Economy,Trade and In-formation Commission of Shenzhen Municipality(Grant No.201901161512),and Guangdong Provincial Key Laboratory(Grant No.2019B121203002).
文摘Optical neural network(ONNs)are emerging as attractive propos-als for machine-learning applications.However,the stability of ONNs decreases with the circuit depth,limiting the scalability of ONNs for practical uses.Here we demonstrate how to compress the circuit depth to scale only logarithmically in terms of the dimension of the data,leading to an exponential gain in terms of noise robustness.Our low-depth(LD)-ONN is based on an architecture,called Optical Com-puTing Of dot-Product UnitS(OCTOPUS),which can also be applied individually as a linear perceptron for solving classification problems.We present both numerical and theoretical evidence showing that LD-ONN can exhibit a significant improvement on robustness,compared with previous ONN proposals based on singular-value decomposition.
基金supported by the Innovate UK Smart (Grant No. 10043476)support from the Royal Commission for the Exhibition of 1851 Research Fellowship。
文摘Optics is an exciting route for the next generation of computing hardware for machine learning,promising several orders of magnitude enhancement in both computational speed and energy efficiency.However, reaching the full capacity of an optical neural network(NN) necessitates that the computing be implemented optically not only for inference but also for training. The primary algorithm for network training is backpropagation, in which the calculation is performed in the order opposite to the information flow for inference. Although straightforward in a digital computer, the optical implementation of backpropagation has remained elusive, particularly because of the conflicting requirements for the optical element that implements the nonlinear activation function. We address this challenge for the first time, we believe, with a surprisingly simple scheme, employing saturable absorbers for the role of activation units. Our approach is adaptable to various analog platforms and materials and demonstrates the possibility of constructing NNs entirely reliant on analog optical processes for both training and inference tasks.
基金This research was supported in part by National Natural Science Foundation of China(61675056 and 61875048).
文摘Optical deep learning based on diffractive optical elements offers unique advantages for parallel processing,computational speed,and power efficiency.One landmark method is the diffractive deep neural network(D^(2) NN)based on three-dimensional printing technology operated in the terahertz spectral range.Since the terahertz bandwidth involves limited interparticle coupling and material losses,this paper extends D^(2) NN to visible wavelengths.A general theory including a revised formula is proposed to solve any contradictions between wavelength,neuron size,and fabrication limitations.A novel visible light D^(2) NN classifier is used to recognize unchanged targets(handwritten digits ranging from 0 to 9)and targets that have been changed(i.e.,targets that have been covered or altered)at a visible wavelength of 632.8 nm.The obtained experimental classification accuracy(84%)and numerical classification accuracy(91.57%)quantify the match between the theoretical design and fabricated system performance.The presented framework can be used to apply a D^(2) NN to various practical applications and design other new applications.
基金This work was supported in part by the National Outstanding Youth Science Fund Project of National Natural Science Foundation of China(62022062)the National Natural Science Foundation of China(61974177,61674119)the Fundamental Research Funds for the Central Universities.
文摘The explosive growth of data and information has motivated various emerging non-von Neumann computational approaches in the More-than-Moore era.Photonics neuromorphic computing has attracted lots of attention due to the fascinating advantages such as high speed,wide bandwidth,and massive parallelism.Here,we offer a review on the optical neural computing in our research groups at the device and system levels.The photonics neuron and photonics synapse plasticity are presented.In addition,we introduce several optical neural computing architectures and algorithms including photonic spiking neural network,photonic convolutional neural network,photonic matrix computation,photonic reservoir computing,and photonic reinforcement learning.Finally,we summarize the major challenges faced by photonic neuromorphic computing,and propose promising solutions and perspectives.
基金supported by the National Natural Science Foundation of China(62035003 and 62235009).
文摘Artificial intelligence(AI)has taken breathtaking leaps forward in recent years,evolving into a strategic technology for pioneering the future.The growing demand for computing power—especially in demanding inference tasks,exemplified by generative AI models such as ChatGPT—poses challenges for conventional electronic computing systems.Advances in photonics technology have ignited interest in investigating photonic computing as a promising AI computing modality.Through the profound fusion of AI and photonics technologies,intelligent photonics is developing as an emerging interdisciplinary field with significant potential to revolutionize practical applications.Deep learning,as a subset of AI,presents efficient avenues for optimizing photonic design,developing intelligent optical systems,and performing optical data processing and analysis.Employing AI in photonics can empower applications such as smartphone cameras,biomedical microscopy,and virtual and augmented reality displays.Conversely,leveraging photonics-based devices and systems for the physical implementation of neural networks enables high speed and low energy consumption.Applying photonics technology in AI computing is expected to have a transformative impact on diverse fields,including optical communications,automatic driving,and astronomical observation.Here,recent advances in intelligent photonics are presented from the perspective of the synergy between deep learning and metaphotonics,holography,and quantum photonics.This review also spotlights relevant applications and offers insights into challenges and prospects.
基金supported in part by the National Natural Science Foundation of China(Grant Nos.62371095,62201096,62401276)by the Natural Science Research Start-up Foundation of Recruiting Talents of Nanjing University of Posts and Telecommunications(Grant No.NY223161)+4 种基金in part by the Jiangsu Provincial Key Research and Development Program(Grant No.BE2022126)the Key R&D Program of Sichuan Province(Grant No.2022ZHCG0041)the National Key Research and Development Program of China(Grant No.2022YFB3206100)the Natural Science Foundation of Sichuan Province(Grant No.2024NSFSC0509)the China Postdoctoral Science Foundation(Grant Nos.2024T170097,2024M760343).
文摘Optical neural networks(ONNs)offer a promising solution for high-performance,energy-efficient artificial intelligence hardware by leveraging the parallelism and speed of light.However,the large-scale implementation of ONNs remains challenging due to the bulky footprint and complex control of optical synapses.In this work,we propose and simulate a plasmonic polarized synaptic architecture that overcomes the diffraction limit and enables ultra-compact ONNs.By tuning the polarization state of incident light,the optical transmittance through each plasmonic unit can be dynamically adjusted to represent a synaptic weight.Our plasmonic structures,with features as small as 40 nm,operate well below this limit in the visible spectrum(400-750 nm).Compared with diffraction and interference-based circuit designs,our proposed method achieves a substantial reduction in synaptic density by factors of 150000-fold and 1500-fold,respectively.Furthermore,we successfully demonstrate a proof-of-concept plasmonic ONN applied to the Canadian Institute for Advanced Research—10 classes(CIFAR-10)dataset using a Visual Geometry Group network with 16 layers(VGG16)model.After training for 80 epochs,the network achieves an accuracy of 93%.The polarization-tunable plasmonics paves the way towards scalable ONNs for next-generation artificial intelligence(AI)accelerators and smart sensors.
基金The authors acknowledge the funding provided by the National Key R&D Program of China(2021YFA1401200)Beijing Outstanding Young Scientist Program(BJJWZYJH01201910007022)+2 种基金National Natural Science Foundation of China(No.U21A20140,No.92050117,No.62005017)programBeijing Municipal Science&Technology Commission,Administrative Commission of Zhongguancun Science Park(No.Z211100004821009)This work was supported by the Synergetic Extreme Condition User Facility(SECUF).
文摘Optical neural networks have significant advantages in terms of power consumption,parallelism,and high computing speed,which has intrigued extensive attention in both academic and engineering communities.It has been considered as one of the powerful tools in promoting the fields of imaging processing and object recognition.However,the existing optical system architecture cannot be reconstructed to the realization of multi-functional artificial intelligence systems simultaneously.To push the development of this issue,we propose the pluggable diffractive neural networks(P-DNN),a general paradigm resorting to the cascaded metasurfaces,which can be applied to recognize various tasks by switching internal plug-ins.As the proof-of-principle,the recognition functions of six types of handwritten digits and six types of fashions are numerical simulated and experimental demonstrated at near-infrared regimes.Encouragingly,the proposed paradigm not only improves the flexibility of the optical neural networks but paves the new route for achieving high-speed,low-power and versatile artificial intelligence systems.
基金National Key Research and Development Program of China(2021YFA1401100)Innovation Group Project of Sichuan Province(20CXTD0090)。
文摘The rapid growth of deep learning applications has sparked a revolution in computing paradigms,with optical neural networks(ONNs)emerging as a promising platform for achieving ultra-high computing power and energy efficiency.Despite great progress in analog optical computing,the lack of scalable optical nonlinearities and losses in photonic devices pose considerable challenges for power levels,energy efficiency,and signal latency.Here,we report an end-to-end all-optical nonlinear activator that utilizes the energy conversion of Brillouin scattering to perform efficient nonlinear processing.The activator exhibits an ultra-low activation threshold(24 nW),a wide transmission bandwidth(over 40 GHz),strong robustness,and high energy transfer efficiency.These advantages provide a feasible solution to overcome the existing bottlenecks in ONNs.As a proof-of-concept,a series of tasks is designed to validate the capability of the proposed activator as an activation unit for ONNs.Simulations show that the experiment-based nonlinear model outperforms classical activation functions in classification(97.64%accuracy for MNIST and 87.84%for Fashion-MNIST)and regression(with a symbol error rate as low as 0%)tasks.This work provides valuable insights into the innovative design of all-optical neural networks.
