189 results about "Optical computing" patented technology

The present disclosure relates to the technical field of optical computing, and particularly relates to an intelligent optical computing chip cluster architecture and system. The architecture comprises: a plurality of optical computing chips, and optical path channels are established between the plurality of optical computing chips by using spatial phase adjustment to modulate light beams; and high-speed optical communication and optical diffraction network computing are performed between the plurality of optical computing chips by performing time-domain intensity modulation on the optical path channels. The present disclosure adopting the above scheme can realize the interconnection computing of optical computing chips in the spatial dimension, and does not rely on electricity for information transmission.

The present invention provides a method for optimizing a neural network model in optical computing. The method comprises: acquiring an initial quantization scaling coefficient and an optical computing system noise suppression coefficient; adjusting the initial quantization coefficient on the basis of the noise suppression coefficient; and then using the adjusted coefficient to quantize a neural network model. The present invention improves the stability and reliability of a model in an optical computing system, thereby reducing the impact of noise of the optical computing system on a neural network model.

An optical computing device includes an optical computing section including a first transmissive optical modulation element having cells each having an independently set phase modulation amount. The first transmissive optical modulation element emits a signal light, in a signal direction, generated when transmitted light beams phase-modulated by the cells interfere with each other and a noise light, in a noise direction, transmitted through the first transmissive optical modulation element without being phase-modulated by the cells. The optical computing device includes an optical sensor that detects the signal light outputted from the optical computing section and generates an electrical signal indicating a result of the detection. The first transmissive optical modulation element is configured such that an impact that the noise light has on the electrical signal is smaller than the impact in a state where the signal direction aligns with the noise direction.

The invention relates to the field of optical computing, and provides a signal processing method and device based on an optical computing chip and a photoelectric fusion computing system.The signal processing device comprises an optical input modulation module, a first weight mapping module, a second weight mapping module and an optical output combining module, and the optical input modulation module converts an input signal into a modulated optical signal; the first weight mapping module obtains a first output optical signal based on the first weight; the second weight mapping module acquires a second output optical signal based on a second weight; the optical output combining module obtains a comprehensive output signal based on the first output optical signal and the second output optical signal; the second weight mapping module comprises a light dimension reduction unit, a low-dimensional light modulation unit and a light dimension raising unit, and the light dimension reduction unit performs low-dimensional projection on a modulated light signal; the low-dimensional optical modulation unit adjusts the second weight based on the second weight update signal and obtains an optical signal after low-dimensional processing; and the light dimension raising unit carries out high-dimensional projection on the light signal after low-dimensional processing.

The invention discloses an RRAM optical computing array reading circuit and method, and the circuit comprises an RRAM optical computing array, and a source line of each unit in the RRAM optical computing array is used for inputting a pulse signal to apply voltage bias; the pulse signal generation unit is used for generating a word line control signal and an RRAM optical calculation array unit bias voltage; the analog multiplexing control unit is used for selecting each subunit of the RRAM optical computing array; the pre-amplification and signal conversion unit is provided with a plurality of circuits and is used for detecting current changes of the photoelectric detection array subunits, converting the current changes into 0 / 1 signals and outputting the 0 / 1 signals to the FPGA control unit; the FPGA control unit is a system microcontroller and is used for controlling the pulse signal generation unit and the output analog multiplexer and processing signals of the pre-amplification and signal conversion unit. According to the invention, the circuit conversion problem caused by a large-span conductive state is solved, the circuit area is simplified, the system power consumption is reduced, and the applicability is high.

The invention discloses an optical calculation chip matrix weight control precision measurement and calibration device and method, and the device mainly comprises a photoelectric detector array module which is used for carrying out the photoelectric signal conversion of an optical signal outputted by an optical calculation chip, a parameter extraction and analysis module which is used for collecting an electric signal, and a calibration module. Calculating the matrix weight control precision and the control rate of the optical calculation chip; and the environment compensation and closed-loop control module is used for monitoring the environment temperature of the optical computing chip and performing compensation calibration control on the bias voltage of the optical computing chip based on the change of the temperature so as to realize accurate regulation and control on the matrix weight. According to the device and the method, the matrix weight control precision and the control rate can be accurately measured, external environment change parameters are accurately sensed, voltage parameters in weight control are changed through a compensation algorithm, and high performance and reliable operation of an optical calculation chip are effectively guaranteed.

The application discloses a kind of diffractive light neural network computing systems and methods for realizing pure optical nonlinearity, belong to photonic computing and artificial intelligence hardware technical field.Its system includes coherent light source, linear light computing unit, nonlinear light activation unit and optical detection unit arranged in order along optical path, wherein linear light computing unit is programmable phase modulation using spatial light modulator, nonlinear light activation unit is programmable binary amplitude modulation using digital micromirror device, and both are directly coupled to form pure optical computing path without photoelectric conversion on optical path.Its method trains neural network through customized loss function, drives nonlinear activation function output to binary convergence, and maps the parameters obtained by training into phase map and binary mask respectively and loads to hardware.The application realizes true all-optical nonlinear computation, with the advantages of high energy efficiency, extremely low delay, compact structure and strong parallel processing capability.

The invention belongs to the technical field of optical calculation, and particularly relates to an intensity-phase mixed silicon-based photon modulator. According to the hybrid silicon-based photon modulator, a PIN micro heater formed by ion implantation is integrated on two arms of a Mach-Zehnder interferometer or a micro-ring resonator, and a phase change material and a protective layer are deposited in a window above the micro heater; electric pulses are applied to metal electrodes at the two ends of the micro heater to drive local reversible crystalline and amorphous transformation of the phase change material, and phase modulation of passing optical signals is achieved; meanwhile, constant voltage signals are applied to the two ends of the PIN micro heater, and intensity modulation is generated on passing optical signals through the carrier absorption effect. The two modulation modes are mutually independent and do not interfere with each other, and two groups of independently variable weight values can be stored in the device. The hybrid silicon-based photon modulator can be used in a large photon neural network, high-speed and high-parallelism convolution operation is carried out, and the overall end-to-end insertion loss of a chip is reduced.

An optical computing system includes: an intensity modulation device group including at least two intensity modulation devices, each of which includes modulation cells, wherein each of the modulation cells of each of the intensity modulation devices carries out intensity modulation with respect to carrier light in accordance with one of signals to generate a signal light beam, and each of the signals corresponds to each of the intensity modulation devices; and a light diffraction element including diffraction cells having respective thicknesses or refractive indices set independently of each other, wherein each of the diffraction cells receives the signal light beam from each of the modulation cells of each of the intensity modulation devices corresponding to each of the diffraction cells, and by causing signal light beams to have respective different optical path lengths to the light diffraction element, the signal light beams have respective different phases.

This application provides a resonant modulator applicable to the field of optoelectronic information technology. The resonant modulator includes a substrate layer, a buried layer, a core layer, and an upper cladding layer. The core layer includes a multimode interference coupler and a connecting waveguide. The multimode interference coupler includes a multimode interference region and a main input port waveguide, a main output port waveguide, a feedback output port waveguide, and a feedback input port waveguide. The two ends of the connecting waveguide are optically connected to the feedback output port waveguide and the feedback input port waveguide, forming an optical feedback loop with the multimode interference coupler. An electro-optic modulation unit is associated with a first region of the connecting waveguide and is used to adjust the effective refractive index of the connecting waveguide by applying an electrical signal to change the resonant state of the resonant modulator. Based on the above structure, the sensitivity to coupling spacing and manufacturing errors can be reduced, and the device operating tolerance and modulation stability can be improved. This application also provides an optical computing system and a photonic integrated system.

This disclosure relates to the field of optical computing technology, and in particular to an intelligent optical computing nonlinear convolutional chip system, method, and architecture, comprising: an input spatiotemporal transformation module for processing an input one-dimensional matrix to generate a complex-weighted extended channel matrix; an input temporal encoding and nonlinear activation module for performing nonlinear activation and first information delay on the complex-weighted extended channel matrix to obtain a matrix; a one-dimensional convolution processing module for performing matrix multiplication on the first matrix to obtain a second matrix; an output temporal encoding and nonlinear activation module for performing nonlinear activation and second information delay on the second matrix to obtain a third matrix; and an output spatiotemporal transformation module for performing complex-weighted channel fusion operation on the third matrix to obtain an output one-dimensional matrix. This disclosure avoids the use of photoelectric conversion and off-chip devices, realizing an all-optical nonlinear convolutional system that can be cascaded to form a large-scale deep all-optical network with high integration, high speed, and strong scalability.

An optical computing device and system, the optical computing device comprising an input module, a feedback module, a reservoir module and a detection module. The input module is configured to compute a first set of optical signals based on a first weight to obtain a second set of optical signals, the first set of optical signals carrying first data to be computed. The feedback module is configured to load the first computation result on feedback light to obtain a first set of feedback optical signals, the first computation result being a computation result of second data by the optical computing device, the second data being data processed before the first data. The reservoir module is configured to compute the first set of feedback optical signals and the second set of optical signals based on a second weight to obtain a third set of optical signals. The detection module is configured to detect optical intensity of the third set of optical signals to obtain a second computation result.

The application provides a photonic neural network system based on a multi-task neural network, which comprises: an optical computing module, taking a photonic artificial intelligence chip as a core, modulating an optical signal, and completing multiplication, summation and nonlinear operation of high-precision and high-speed optical analog quantities; a multi-task neural network module, having a main branch neural network and multiple secondary branch neural networks, and capable of simultaneously processing multiple classification tasks and regression tasks; and an optical computing communication and control module based on FPGA, connected with the optical computing module and the multi-task neural network module, and used for realizing real-time and high-speed communication of the optical computing module and the multi-task neural network module. The system has the characteristics of high bandwidth and low energy consumption, utilizes high-dimension parallel computing to greatly reduce the time of original neural network serial operation, has the advantages of large network scale and fast parallel operation speed, and can improve recognition accuracy and shorten training time on the basis of a single task.

The application provides an optical signal synchronous calibration device and method of an optical computing system, the device comprising: an optical probe array arranged on a chip, an AI coprocessor, an optical path tuning engine and a low-speed electrical feedback line arranged on a board card; the optical probe array detects optical signals of wavelength channels corresponding to a plurality of optical computing units and converts the optical signals into electrical signals; the AI coprocessor drives the corresponding optical computing unit to emit a detection pulse according to a pre-allocated time slot index variable, extracts a phase error from the electrical signal and calculates a final calibration quantity based on the phase error; the optical path tuning engine performs an optical path tuning operation based on the final calibration quantity to calibrate the error of the corresponding optical computing unit; and the low-speed electrical feedback line transmits the electrical signal to the AI coprocessor and transmits the detection pulse to the chip. The application realizes real-time calibration in a dynamic environment through spatial distributed detection and board-level centralized processing cooperation.

The application provides an FPGA-based optical computing communication and control module and belongs to the technical field of optical computing.The development board comprising an FPGA chip, an AD / DA module, an output input interface module, a clock module and a memory module is adopted, so that the module can perform four-channel AD / DA signal processing and has rich interfaces to support high-speed information processing and transmission, thereby providing an ideal communication platform for the optical computing module and the neural network and providing the ability of real-time high-speed weight updating of the optical computing module.

The invention is suitable for the technical field of optical computing, and provides a knowledge distillation-based on-chip optical neural network design method, which comprises the following steps of: constructing an initial population of an optimization algorithm based on a preset process constraint condition; the plurality of individuals in the initial population are in one-to-one correspondence with the plurality of groups of optical neural network parameters; performing iterative optimization on the initial population by using an optimization algorithm to obtain an optimal optical neural network parameter meeting a preset process constraint condition; constructing an optical neural network based on the optimal optical neural network parameters, and taking the constructed optical neural network as an on-chip optical neural network; the calculation process of the fitness value of the individual in the optimization algorithm comprises the following steps: training the optical neural network corresponding to the individual by using the knowledge distillation framework, and determining the fitness value based on the training loss value and the physical parameter of the optical neural network. According to the invention, the area of the on-chip optical neural network is reduced, the complexity of the on-chip optical neural network is reduced, and the performance of the on-chip optical neural network in processing complex tasks is improved.

The invention discloses a multi-wavelength channel diffraction neural network system based on threshold screening and a processing method, and belongs to the technical field of artificial intelligence and optical computing. Comprising an encoding module used for encoding task information corresponding to different to-be-processed tasks to multiple light beams with different working wavelengths output by a light source module to form multiple task channel light fields; the optical diffraction module comprises a plurality of modulation layers; the optical diffraction module is used for receiving the task channel light fields and realizing diffraction modulation and threshold screening of the task channel light fields by changing the metasurface structures on different modulation layers; each task channel light field shares the same set of diffraction modulation parameters, namely parameters such as phase, mask generation and threshold; threshold screening: comparing the mask generation parameter with a threshold parameter, performing binarization selection on the phase parameter according to a comparison result, and screening out a target phase parameter; the output detection module detects the output light intensity. Task channels are expanded, and the system integration degree is improved.

The application relates to the field of ultra-wideband radar imaging and optical computing technology, in particular to a light computing real-time processing method for ultra-wideband radar imaging. The method comprises core steps of signal receiving and electro-optical modulation, optical pulse compression, initial phase compensation based on Kalman filtering, envelope alignment, coherence recovery and two-dimensional grid projection. Based on the ultrafast linear processing advantage and algorithm optimization principle of light computing, the application utilizes optical finite impulse response filters and adjustable delay line technology to disassemble and reconstruct the traditional imaging method, and reduces the end-to-end processing delay to the sub-millisecond level. The application breaks through the real-time processing bottleneck of ultra-wideband radar imaging, provides effective theoretical support and method basis for developing a full real-time ultra-wideband two-dimensional fine imaging system, and is widely applicable to application fields with high real-time requirements such as military reconnaissance and geological exploration.

We propose a hybrid wave-based logic model for optical computing that integrates both Total Wave Modified Schrödinger Equation (TWMSE) formulations and non-TWMSE interference field logic. This system enables deterministic collapse behavior through physical wave interactions, rather than relying on conventional transistor logic or probabilistic quantum models. In the TWMSE regime, collapse thresholds are governed by field interference between signal and control waves, allowing dynamic selection of tasks based on intensity and phase alignment. In the generalized non-TWMSE regime, similar collapse-like logic can be achieved using thresholded interference models that are classically engineered for optical domain systems. Together, these formulations enable adaptive task selection, nonlinear logic gating, and scalable parallel processing in optical chips. Our architecture supports real-time wavefield decisions and is compatible with standard photonic components such as interferometers, microring resonators, variable attenuators, and optical threshold comparators. The result is a unified logic framework for next-generation AI hardware that harnesses both quantum-inspired and classical wave interference dynamics.

The purpose of the present invention is to provide an optical computer having high scalability and capable of reducing noise. This optical computer is characterized by including: an input layer which accepts an input signal as a computer; a reservoir layer which responds to a signal input from the input layer, and in which using a plurality of degenerate optical parametric oscillator (DOPO) pulses of a DOPO that are generated by an optical phase-sensitive amplifier (PSA) based on pump light as nodes, a connection relation between the nodes is determined by a measurement feedback method or a method equivalent to the measurement feedback method; and an output layer which outputs an output signal based on the respective responses of the plurality of DOPO pulses to the input signal from the reservoir layer.

The application belongs to the technical field of semiconductor optoelectronic devices, and particularly relates to a light computing array based on a reverse bias germanium modulator, which comprises a cross matrix formed by multiple rows of parallel input light waveguides and multiple columns of parallel output light waveguides; an optical detector is arranged at the output end of each column of the output light waveguides; multiple light splitters; multiple germanium modulators, each corresponding to one cross waveguide, and the first input end of the germanium modulator being connected with the output end of the light splitter through a transmission light waveguide; and multiple light combiners, each being connected with the output end of the germanium modulator and the output light waveguide. Through the above structure, the application comprehensively provides a high-response computing array with a thermal compensation mechanism and capable of reducing light reflection, so as to realize a scheme of a large-scale integrated computing array with low power consumption under the condition of meeting frequent switching.

The invention discloses an optical switch device and a complex-valued optical neural network system, and belongs to the technical field of optical computing. The optical switch device is a two-input two-output device and comprises first to sixth waveguides, and first and second multimode interference couplers. Phase shifters are attached to the third waveguide and the fifth waveguide and / or the fourth waveguide and the sixth waveguide to form a group of configuration units which are used for loading real part and imaginary part data of a complex number. The phase shifter comprises a nonvolatile phase change material layer and a heating layer which are sequentially arranged from bottom to top, is used for realizing the function of the phase shifter, greatly reduces the static power consumption, improves the energy efficiency, is suitable for realizing the complex value optical neural network, and is also beneficial to large-scale integration. The complex-valued optical neural network system constructed based on the optical switch integrates input, calculation, reference and coherent detection modules, can synchronously complete amplitude and phase detection of optical signals, effectively supports complex-valued optical calculation, and has the advantages of low power consumption and high integration level.

The application discloses an optical matrix vector multiplier based on a unitary matrix-diagonal matrix decomposition. The multiplier comprises a Mach-Zehnder interferometer array for realizing a synthetic unitary matrix Omega and a modulation unit for realizing a corresponding diagonal matrix Sigma, wherein a target weight matrix M satisfies a decomposition form of M=ΩΣ. The method fuses two unitary matrices required for mapping in a traditional singular value decomposition into a single synthetic unitary matrix Omega, so that a unitary transformation can be completed by only one interferometer array of optical hardware, and a complete matrix multiplication is realized in cooperation with the modulation unit. Through algorithm-hardware collaborative design, the number of basic optical interference units required is significantly reduced, the hardware complexity, manufacturing cost and system power consumption of the optical neural network are greatly reduced while the computing performance is maintained, and an effective solution is provided for efficient optical computing acceleration of artificial intelligence.

The application relates to the technical field of optical computing and artificial intelligence, and provides a multi-task diffraction neural network device and processing method based on multi-wavelength parallelism, the device comprising an input unit, a diffraction modulation structure, a light detection assembly and a processing unit; the input unit is used for modulating the inputs of N tasks to N wavelengths, inputting a mixed light beam formed after light field superposition into the diffraction modulation structure, and the diffraction modulation structure is used for outputting after parallel processing of each wavelength component of the input mixed light beam; the light detection assembly is used for detecting the light intensity of the output plane of the diffraction modulation structure; the output plane comprises M categories of detection regions, and each detection region comprises N sub-regions; and the processing unit is used for determining the inference results of the N tasks corresponding to the N wavelengths according to the light intensity distribution of the corresponding sub-region in each detection region. The problems that D2NN can only adapt to a single deep learning task and has poor universality are solved, multi-task parallelism is realized, and the universality is improved.

The invention provides an optical processing chip and an optical computing system, and relates to the technical field of photoelectronics. The optical coupling structure comprises an optical coupling structure, a micro-nano superstructure and an optical signal receiving structure. The first surface of the optical coupling structure is provided with a plurality of optical output ports used for outputting a plurality of first optical signals. The micro-nano superstructure is arranged on the first surface and comprises a plurality of superstructure units, and orthographic projections of the superstructure units on the first surface cover the light output ports in a one-to-one correspondence mode. Each super-structure unit comprises a plurality of nano-columns which are arranged at intervals and used for processing a plurality of first optical signals into second optical signals, and the first optical signals are different from the second optical signals. And the optical signal receiving structure is connected with the micro-nano metamaterial structure, and the optical signal receiving structure is configured to receive the second optical signal and obtain information carried by the second optical signal. Parallel computing is realized by using a plurality of superstructure units, so that the computing efficiency of the optical processing chip is improved, and meanwhile, the integration level of the optical processing chip is enhanced.

The application discloses an end-to-end optical computing chip based on multi-modal analog signal fusion processing, and belongs to the technical field of optical computing. The chip comprises a multi-modal input fusion front-end module, a fiber input interface and an end-to-end inference module. The multi-modal input fusion front-end module can convert different types of original analog signals such as images, spectra and radio frequencies into unified broadband spectral input signals. The end-to-end inference module builds a deep optical neural network, which comprises a "sensing-convolution integrated" unit realized by an arrayed waveguide grating (AWG). The end-to-end inference module also comprises an optoelectronic nonlinear-pooling integrated unit, which simultaneously realizes the functions of average pooling and nonlinear activation, and effectively compensates for optical path loss through injection of a light source. Finally, the signals processed by multiple layers are integrated by a full connection layer, and a classification result is output by an output layer. The chip architecture realizes direct and efficient processing of multi-modal analog signals.

The application provides a target detection method and system based on photoelectric hybrid computing. The method comprises the following steps: acquiring an image, and pre-processing the image; inputting the image into a light computing unit to perform linear change calculation, and outputting optical characteristics of the image; the light computing unit comprises a Mach-Zehnder interferometer array; inputting the optical characteristics into an electronic computing unit to perform nonlinear operation target detection, and outputting a target detection result of the image; the electronic computing unit comprises an electrical target detection model constructed based on a single multi-frame detector. In this way, efficient target detection can be realized through deep fusion of optical computing and electronic computing, the computing efficiency is optimized, the energy efficiency ratio is improved, the delay is reduced, and the system flexibility is enhanced.

This invention relates to the field of optical-electric hybrid packaging technology, specifically to an optoelectronic hybrid packaging system and computing method. The 3D packaging system enables large-scale high-speed computation. The 3D packaging system includes: a printed circuit board, a metal frame, and an optical computing layer arranged sequentially. The optical computing layer includes: an optical computing region comprising: a first optical chip with its first optical port adjacent to a first side area for connection to an optical fiber; a second optical chip optically connected to the first optical chip; and an electrical transmission layer with an electrical transmission region corresponding to the optical computing region, comprising: multiple electrical chips; each electrical chip having a first electrical channel and a second electrical channel, the first electrical channel connected to a converter, and the second electrical channel connected to an adjacent electrical chip. This invention provides a 3D packaging system for the large-scale joint packaging of multiple optical and electrical chips, thereby facilitating the realization of ultra-large-scale optical computation.

Embodiments of the present application disclose an optical computing device and method, relating to the field of information technology, to provide low computing latency optical computing. The device comprises an optical analog domain module, a first optoelectronic conversion module and an electrical analog domain module. The optical analog domain module is configured to receive a first electrical signal and an optical signal, modulate the received optical signal according to the first electrical signal to generate a first optical signal, and send the first optical signal to the first conversion module. The first optoelectronic conversion module is configured to convert the first optical signal into a first electrical analog signal, send the first electrical analog signal to the electrical analog domain module, convert the rearranged first electrical analog signal into a second optical signal, and send the second optical signal to the optical analog domain module. The electrical analog domain module is configured to rearrange the first electrical analog signal, and send the rearranged first electrical analog signal to the first optoelectronic conversion module. The first electrical signal indicates a target weight, and the first optical signal indicates an intermediate calculation result of to-be-calculated data and the target weight.