Reconfigurable diffraction universal intelligent light calculation method, architecture and system

Abstract

The invention relates to the technical field of optical computing, in particular to a reconfigurable diffraction universal intelligent optical computing method, architecture and system. The method comprises the following steps: obtaining a first matrix corresponding to an input mode, and obtaining a signal parameter matrix through a modulator array; obtaining a second matrix of the input parameters, and processing the second matrix through a modulation core to obtain a corresponding modulation parameter matrix; randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; inputting the fusion matrix into a target diffraction propagation model to obtain a diffraction transmission matrix; and outputting a target form signal based on the diffraction transmission matrix. According to the invention, through dynamic adjustment of the modulation parameters and random fusion of the modulation parameter matrix and the signal parameter matrix, any reconfigurable characteristic of the diffraction calculation result is realized, and the diffraction calculation and the reconstruction modulation parameters are separated in space, so that the diffraction calculation has general calculation capability, and the calculation efficiency is improved. And the challenge that the diffraction network is difficult to reconstruct is broken through.

Description

Technical Field

[0001] The present disclosure relates to the field of optical computing technology, and in particular to a reconfigurable diffraction universal intelligent optical computing method, architecture, and system. Background Art

[0002] With the rapid development of artificial intelligence and scientific computing, the complexity and scale of computing needs are increasing. However, existing microelectronic computing chips face performance bottlenecks (such as speed and energy consumption) when handling rapidly growing computing power demands, making it difficult to effectively meet the increasingly stringent computing power and power consumption requirements of large-scale, complex algorithms. Optical computing, as a new computing paradigm, leverages the propagation characteristics and parallel processing capabilities of light, demonstrating the advantages of high computing power and low energy consumption. This makes optical computing technology considered the key to breaking through existing computing bottlenecks. Summary of the Invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the first objective of the present disclosure is to propose a reconfigurable diffraction universal intelligent optical computing method.

[0005] The second objective of the present disclosure is to provide a reconfigurable diffraction universal intelligent optical computing system.

[0006] To achieve the above objectives, the first embodiment of the present disclosure proposes a reconfigurable diffraction universal intelligent optical computing method, the method comprising: Obtain the first matrix corresponding to the input mode, and obtain the signal parameter matrix through the modulator array; Obtaining a second matrix of input parameters, and processing the second matrix using a modulation kernel to obtain a corresponding modulation parameter matrix; Randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; Inputting the fusion matrix into the target diffraction propagation model to obtain a diffraction transmission matrix; Based on the diffraction transfer matrix, a target form signal is output.

[0007] Optionally, the processing the second matrix by using a modulation kernel to obtain a corresponding modulation parameter matrix includes: Determine the parameter type corresponding to the input parameter; Determining a target modulation parameter corresponding to the modulation core based on the parameter type; Based on the target modulation parameters, the second matrix is ​​processed by a modulation kernel to obtain a corresponding modulation parameter matrix.

[0008] Optionally, randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fused matrix includes: fusing the modulation parameter matrix and the signal parameter matrix in parallel or cross-fusing to obtain a fused matrix.

[0009] Optionally, inputting the fusion matrix into the target diffraction propagation model to obtain the diffraction transfer matrix includes: inputting the fusion matrix into the target diffraction propagation model, and obtaining the diffraction transfer matrix through the target diffraction kernel in the target diffraction propagation model, wherein the target diffraction kernel is trained.

[0010] Optionally, outputting a target form signal based on the diffraction transmission matrix includes: determining whether the diffraction transfer matrix satisfies an output condition; If the diffraction transfer matrix meets the output conditions, determining the target form to be output; Determining a corresponding signal feed technology based on the target form; Based on the signal feeding technology and the diffraction transmission matrix, a target form signal is output.

[0011] Optionally, the method further includes: If the diffraction transfer matrix does not meet the output condition, adjusting the parameters of the modulation kernel based on the diffraction transfer matrix; Based on the diffraction transfer matrix, the above steps are repeated until the diffraction transfer matrix meets the output conditions and the target form signal is output.

[0012] To achieve the above objectives, a second embodiment of the present disclosure provides a reconfigurable diffraction universal intelligent optical computing system, the system comprising: A signal feeding module is used to obtain a first matrix corresponding to the input mode and obtain a signal parameter matrix through a modulator array; a parameter modulation module, configured to obtain a second matrix of input parameters and process the second matrix using a modulation kernel to obtain a corresponding modulation parameter matrix; a channel synthesis module, configured to randomly fuse the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; a diffraction propagation module, configured to input the fusion matrix into a target diffraction propagation model to obtain a first diffraction transmission matrix; The result output module is used to output a target form signal based on the first diffraction transmission matrix.

[0013] Another object of the present disclosure is to propose a reconfigurable diffraction universal intelligent optical computing architecture, comprising: at least one reconfigurable diffraction universal intelligent optical computing system.

[0014] In summary, the reconfigurable diffraction universal intelligent optical computing method and system provided by the present disclosure realizes the arbitrary reconfigurability of the diffraction calculation results through the dynamic adjustment of the modulation parameters and the random fusion of the modulation parameter matrix and the signal parameter matrix. In addition, by spatially separating the diffraction calculation and the reconstruction of the modulation parameters, the diffraction calculation is endowed with universal computing capabilities, thus breaking through the challenge of difficult diffraction network reconstruction.

[0015] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and / or additional aspects and advantages of the present disclosure will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A schematic diagram of a flow chart of a reconfigurable diffraction universal intelligent light calculation method provided by an embodiment of the present disclosure; Figure 2 A schematic diagram of a reconfigurable diffraction universal intelligent light computing method provided by an embodiment of the present disclosure; Figure 3 A schematic diagram of a reconfigurable diffraction universal intelligent light computing method provided by an embodiment of the present disclosure; Figure 4 A schematic diagram of the structure of a reconfigurable diffraction universal intelligent optical computing system provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0017] The following describes in detail embodiments of the present disclosure, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0018] Currently, diffraction neural networks offer the advantages of ultra-high computing power and ultra-low energy consumption by combining neural networks with ultra-high-density optical diffraction. However, existing diffraction optical computing chips cannot be reconfigured for different application scenarios and computing requirements, resulting in a lack of structural flexibility. Consequently, the difficulty of reconfiguring diffraction calculations leads to the limited computational functionality of diffraction neural networks, significantly limiting their application scenarios and practical value.

[0019] Furthermore, after the diffraction network chip is fabricated, its structure or optical properties cannot be arbitrarily changed, resulting in a very simple architecture and functionality of the chip, making it difficult to effectively meet the needs of multi-function and multi-scenario applications, greatly restricting the actual application value of the chip. Existing technologies have explored and optimized on-chip reconfigurable technologies at the photonic device level (such as photonic devices based on waveguide structures), such as non-volatile control based on phase change materials, transient control technology based on plasma dispersion effects and thermo-optical effects, etc. In existing technologies, for laser direct writing all-optical reconstruction technology, the size mismatch between neurons and the reconstruction method (laser spot) leads to reconstruction crosstalk problems between high-density neurons. The reconstruction process is highly dependent on the microscopic imaging system and high-precision mechanical displacement system, which determines the dual limitations of reconstruction scale and reconstruction efficiency. For example, for electro-optical control technology based on plasma dispersion and thermo-optical effects, the reconstruction effect (such as resolution and scale) is heavily dependent on the electrode size and complex manufacturing process, which directly weakens its application value. Based on this, the current on-chip reconstruction technology cannot effectively solve the crosstalk problem and low efficiency of large-scale reconstruction caused by high-density reconstruction of subwavelength-scale neurons. It is particularly difficult to accurately reconstruct subwavelength-scale diffraction neurons.

[0020] The present disclosure is described in detail below with reference to specific embodiments.

[0021] Figure 1 This is a flow chart of a reconfigurable diffraction universal intelligent light calculation method provided by the embodiment of the present disclosure. Figure 1 As shown, the reconfigurable diffraction universal intelligent light calculation method may include the following steps: Step 101: Obtain a first matrix corresponding to an input mode, and obtain a signal parameter matrix through a modulator array; Step 102: Obtain a second matrix of input parameters, and process the second matrix using a modulation kernel to obtain a corresponding modulation parameter matrix; Step 103: randomly fuse the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; Step 104: Input the fusion matrix into the target diffraction propagation model to obtain a diffraction transmission matrix; Step 105: Output a target form signal based on the diffraction transmission matrix.

[0022] In one embodiment of the present disclosure, the above-mentioned reconfigurable diffraction universal intelligent light computing method is applicable to different tasks, such as image classification and multimodal information recognition.

[0023] In one embodiment of the present disclosure, the input mode can be set as needed, for example, image, text, or audio.

[0024] In one embodiment of the present disclosure, a signal parameter matrix corresponding to the first matrix can be obtained through a modulator array. In one embodiment of the present disclosure, the signal parameter matrix can be an amplitude and / or phase matrix corresponding to the first matrix.

[0025] Furthermore, in one embodiment of the present disclosure, the second matrix of the input parameters may be obtained by processing the required parameters through the corresponding modulator array. In one embodiment of the present disclosure, the required parameters may be voltages.

[0026] Furthermore, in one embodiment of the present disclosure, the method of processing the second matrix using the modulation kernel to obtain the corresponding modulation parameter matrix may include the following steps: Step 1021: Determine the parameter type corresponding to the input parameter; Step 1022: Determine the target modulation parameter corresponding to the modulation core based on the parameter type; Step 1023: Based on the target modulation parameters, the second matrix is ​​processed by the modulation kernel to obtain a corresponding modulation parameter matrix.

[0027] In one embodiment of the present disclosure, the parameter type corresponding to the input parameter can be determined manually or by the type of the input parameter.

[0028] Furthermore, in one embodiment of the present disclosure, the modulation core may be trained, and different parameter types correspond to different target modulation parameters. In one embodiment of the present disclosure, the implementation method of the modulation core may include, but is not limited to, spatial light or on-chip, and may include, but is not limited to, thermal modulators, carrier modulators, or phase change materials, and may include, but is not limited to, specific material platforms such as silicon, silicon dioxide, silicon nitride, lithium niobate, etc.

[0029] Furthermore, in one embodiment of the present disclosure, after the target modulation parameters are obtained through the above steps, the second matrix can be processed by a modulation kernel based on the target modulation parameters to obtain a corresponding modulation parameter matrix.

[0030] In one embodiment of the present disclosure, after obtaining the modulation parameter matrix and the signal parameter matrix through the above steps, the modulation parameter matrix and the signal parameter matrix can be randomly fused to obtain a corresponding fused matrix. In one embodiment of the present disclosure, the modulation parameter matrix and the signal parameter matrix can be randomly fused as needed, thereby achieving arbitrary reconfigurability of the diffraction calculation results.

[0031] Furthermore, in one embodiment of the present disclosure, a method of randomly fusing a modulation parameter matrix and a signal parameter matrix to obtain a fused matrix may include: fusing the modulation parameter matrix and the signal parameter matrix in parallel or cross-fusing to obtain a fused matrix.

[0032] Furthermore, in one embodiment of the present disclosure, after obtaining the fused matrix through the above steps, the fused matrix can be input into the target diffraction propagation model to obtain a diffraction transfer matrix. In one embodiment of the present disclosure, the method of inputting the fused matrix into the target diffraction propagation model to obtain the diffraction transfer matrix can include: inputting the fused matrix into the target diffraction propagation model, and obtaining the diffraction transfer matrix using a target diffraction kernel in the target diffraction propagation model, wherein the target diffraction kernel is trained.

[0033] Furthermore, in one embodiment of the present disclosure, the specific implementation of the diffraction kernel may include but is not limited to single-layer or multi-layer diffraction, and the form of neurons may include but is not limited to neurons of various shapes, such as rectangular, circular, conical, elliptical, etc. In one embodiment of the present disclosure, a corresponding target diffraction kernel may be obtained by training the corresponding diffraction kernel to determine the implementation of the target diffraction kernel and the form of neurons.

[0034] Furthermore, in one embodiment of the present disclosure, after obtaining the diffraction transfer matrix through the above steps, a target form signal can be output based on the diffraction transfer matrix. Specifically, in one embodiment of the present disclosure, the above method of outputting a target form signal based on the diffraction transfer matrix can include the following steps: Step 1051: Determine whether the diffraction transfer matrix meets the output conditions; Step 1052: If the diffraction transfer matrix meets the output conditions, determine the target form to be output; Step 1053: Determine a corresponding signal feed technology based on the target form; Step 1054: Output a target signal based on the signal feed technology and the diffraction transmission matrix.

[0035] In one embodiment of the present disclosure, the output condition may be the number of times of loop calculation.

[0036] Furthermore, in one embodiment of the present disclosure, when the reconfigurable diffraction universal intelligent optical computing method is applied to different tasks, the corresponding output conditions are also different. For example, in one embodiment of the present disclosure, if the reconfigurable diffraction universal intelligent optical computing method is applied to image classification, the corresponding output condition may be a single loop calculation; if the reconfigurable diffraction universal intelligent optical computing method is applied to multimodal information recognition, the corresponding output condition may be a four-loop calculation.

[0037] Furthermore, in one embodiment of the present disclosure, if the diffraction transfer matrix satisfies the output conditions, the target form to be output can be determined. In one embodiment of the present disclosure, the target form to be output can be determined by user input, and the target form includes, but is not limited to, a waveguide signal, an optical radiation signal, and an electrical signal, including, but not limited to, the intensity and spatial distribution of an optical signal, or the voltage and current of an electrical signal.

[0038] Furthermore, in one embodiment of the present disclosure, after determining the target form through the above steps, the corresponding signal feeding technology can be determined according to the target form, so as to output the target form signal based on the signal feeding technology and the diffraction transmission matrix.

[0039] In one embodiment of the present disclosure, different target forms correspond to different signal feeding technologies.

[0040] Specifically, in one embodiment of the present disclosure, the signal feeding technology corresponding to the waveguide signal may be a mode converter based on a waveguide array; the signal feeding technology corresponding to the optical radiation signal may be an optical radiator based on a coupler array; and the signal feeding technology corresponding to the electrical signal may be a photoelectric converter based on a photodetector array.

[0041] Furthermore, in one embodiment of the present disclosure, the above method may further include the following steps: Step 1055: If the diffraction transfer matrix does not meet the output condition, adjust the parameters of the modulation kernel in the parameter modulation module based on the diffraction transfer matrix; Step 1056: Based on the diffraction transfer matrix, repeat the above steps until the diffraction transfer matrix meets the output conditions and outputs the target form signal.

[0042] In one embodiment of the present disclosure, after adjusting the parameters of the modulation kernel in the parameter modulation module based on the diffraction transfer matrix, the diffraction transfer matrix can be used as the first matrix and the second matrix, and the above steps are repeated to obtain an updated diffraction transfer matrix until the updated diffraction transfer matrix meets the output conditions and outputs the target form signal.

[0043] In one embodiment of the present disclosure, the above steps may implement dynamic adjustment of target modulation parameters based on the dynamic modulation characteristics of the modulation core and feedback of the output diffraction transfer matrix.

[0044] Figure 2 and Figure 3 A schematic diagram of a reconfigurable diffraction universal intelligent light computing method provided in an embodiment of the present disclosure.

[0045] like Figure 2As shown in the figure, after the signal parameter matrix and the modulation parameter matrix are randomly fused to obtain a fusion matrix, the fusion matrix is ​​input into the target diffraction propagation model, and the diffraction transmission matrix is ​​obtained through the diffraction kernel in the target diffraction propagation model. The modulation parameters in the modulation kernel are adjusted through the diffraction transmission matrix, and then the above steps are repeated through the diffraction transmission matrix to obtain the target form signal.

[0046] like Figure 3 As shown, after adjusting the modulation parameters in the modulation core through the diffraction transfer matrix, the corresponding modulation core 1, modulation core 2, modulation core 3 until modulation core N can be generated respectively, and the target form signal can be obtained in turn through the corresponding modulation core and diffraction core.

[0047] In summary, the reconfigurable diffraction universal intelligent optical computing method provided in this embodiment realizes the arbitrary reconfiguration characteristics of the diffraction calculation results through dynamic adjustment of modulation parameters and random fusion of the modulation parameter matrix and the signal parameter matrix. In addition, by spatially separating the diffraction calculation and the reconstruction of the modulation parameters, the diffraction calculation is endowed with universal computing capabilities, thus breaking through the challenge of difficult diffraction network reconstruction.

[0048] In order to implement the above embodiment, Figure 4 A reconfigurable diffraction universal intelligent optical computing system provided in an embodiment of the present disclosure includes: The signal feeding module 401 is used to obtain a first matrix corresponding to the input mode and obtain a signal parameter matrix through the modulator array; A parameter modulation module 402 is configured to obtain a second matrix of input parameters and process the second matrix using a modulation kernel to obtain a corresponding modulation parameter matrix; The channel synthesis module 403 is used to randomly fuse the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; The diffraction propagation module 404 is used to input the fusion matrix into the target diffraction propagation model to obtain a first diffraction transmission matrix; The result output module 405 is configured to output a target form signal based on the first diffraction transmission matrix.

[0049] Among them, the signal feed port and parameter modulation port of the diffraction module do not need to have a clear boundary in form. They can be the same or different. Therefore, in network design, their spatial distribution will be included in the training as part of the parameters.

[0050] In one embodiment of the present disclosure, the signal feed module port and the parameter modulation module port of the above-mentioned diffraction propagation module do not need to have obvious boundaries in form, and can be the same or different. Based on this, their spatial distribution will be included in the training as part of the parameters in the network design.

[0051] Furthermore, in one embodiment of the present disclosure, the essence of the on-chip reconstruction technology for optical devices or chips is the interaction between light and matter. The core difference between the currently widely used laser direct writing reconstruction technology, plasma dispersion effect modulation technology, and thermo-optical effect modulation technology lies in the differences in the energy carriers, objects of action, modes of action, and effects of light-matter interactions. Based on this, in one embodiment of the present disclosure, a reconfigurable diffraction universal intelligent optical computing system is proposed, which uses a new modular reconstruction approach. During the reconstruction process, the core focus is on the control effect of external modulation parameters on the diffraction propagation output results, and the predetermined control effect is achieved by optimizing the deployment of control parameters.

[0052] Furthermore, in one embodiment of the present disclosure, the channel boundary ablation for signal input and parameter input is essentially the same. Considering the challenges faced by traditional control technologies for dense diffraction neuron control, the present embodiment proposes a more flexible selection of signal and parameter input channels for the reconfigurable diffraction universal intelligent optical computing system, without focusing on the specific form of control energy. This is expected to provide better compatibility with various current control technologies.

[0053] Optionally, in one embodiment of the present disclosure, the parameter modulation module 402 is specifically configured to: Determine the parameter type corresponding to the input parameter; Based on the parameter type, determining the target modulation parameter corresponding to the modulation core; Based on the target modulation parameters, the second matrix is ​​processed by the modulation kernel to obtain a corresponding modulation parameter matrix.

[0054] Optionally, in one embodiment of the present disclosure, the channel synthesis module 403 is specifically configured to: The modulation parameter matrix and the signal parameter matrix are fused in parallel or cross-fused to obtain a fusion matrix.

[0055] Optionally, in one embodiment of the present disclosure, the diffraction propagation module 404 is specifically configured to: The fusion matrix is ​​input into the target diffraction propagation model, and the diffraction transmission matrix is ​​obtained through the target diffraction kernel in the target diffraction propagation model, wherein the target diffraction kernel is trained.

[0056] Optionally, in one embodiment of the present disclosure, the result output module 405 is specifically configured to: Determine whether the diffraction transfer matrix meets the output conditions; If the diffraction transfer matrix meets the output conditions, determine the target form to be output; Determine the corresponding signal feed technology based on the target form; Based on signal feed technology and diffraction transmission matrix, the target form signal is output.

[0057] Optionally, in one embodiment of the present disclosure, the above device is further used to: If the diffraction transfer matrix does not meet the output condition, the parameters of the modulation kernel are adjusted based on the diffraction transfer matrix; Based on the diffraction transfer matrix, the above steps are repeated until the diffraction transfer matrix meets the output conditions and the target form signal is output.

[0058] The present disclosure also proposes a reconfigurable diffraction universal intelligent optical computing architecture, which may include: at least one of the above-mentioned reconfigurable diffraction universal intelligent optical computing systems, and the reconfigurable diffraction universal intelligent optical computing systems are cascaded to complete corresponding tasks.

[0059] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this disclosure are in compliance with relevant laws and regulations and do not violate public order and good morals.

[0060] It is important to note that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold beyond these legitimate uses. Furthermore, such collection / sharing should be conducted only after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes the relevant user information before using the feature. Furthermore, any necessary steps must be taken to safeguard and secure access to such personal information and ensure that others with access to personal information comply with its privacy policy and procedures.

[0061] This disclosure contemplates providing implementations that allow users to selectively block the use or access of personal information data. Specifically, this disclosure contemplates providing hardware and / or software to prevent or block access to such personal information data. Risks can be minimized by limiting data collection and deleting data once it is no longer needed. Furthermore, where applicable, such personal information can be de-identified to protect user privacy.

[0062] The acquisition, transmission, storage, use, and processing of data in the technical solution disclosed herein are in compliance with the relevant provisions of national laws and regulations.

[0063] It should be noted that in an embodiment of the present disclosure, certain software, components, models, and other existing solutions in the industry may be mentioned. They should be regarded as exemplary and their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0064] In the descriptions of the aforementioned embodiments, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0065] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0066] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0067] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" is any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (not exhaustive) of computer-readable media include: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or otherwise processing it in a suitable manner if necessary, and then storing it in a computer memory.

[0068] It should be understood that various parts of the present disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0069] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above-mentioned embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0070] In addition, the functional units in the various embodiments of the present disclosure may be integrated into a single processing module, each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0071] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present disclosure have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. A person of ordinary skill in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A reconfigurable diffraction universal intelligent optical computing method, characterized in that: The method comprises: Obtain a first matrix corresponding to the input mode, and obtain a signal parameter matrix through the modulator array; Obtaining a second matrix of input parameters, and processing the second matrix through a modulation kernel to obtain a corresponding modulation parameter matrix; Randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; Inputting the fusion matrix into the target diffraction propagation model to obtain a diffraction transmission matrix; Based on the diffraction transmission matrix, a target form signal is output.

2. The method according to claim 1, characterized in that The step of processing the second matrix by using a modulation kernel to obtain a corresponding modulation parameter matrix includes: Determine the parameter type corresponding to the input parameter; Based on the parameter type, determining a target modulation parameter corresponding to the modulation core; Based on the target modulation parameters, the second matrix is ​​processed by a modulation kernel to obtain a corresponding modulation parameter matrix.

3. The method according to claim 1, characterized in that The randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fused matrix includes: fusing the modulation parameter matrix and the signal parameter matrix in parallel or cross-fusing to obtain a fused matrix.

4. The method according to claim 1, characterized in that: The step of inputting the fusion matrix into the target diffraction propagation model to obtain the diffraction transmission matrix includes: inputting the fusion matrix into the target diffraction propagation model, and obtaining the diffraction transmission matrix through the target diffraction kernel in the target diffraction propagation model, wherein the target diffraction kernel is trained.

5. The method according to claim 1, characterized in that The outputting of a target form signal based on the diffraction transmission matrix comprises: Determining whether the diffraction transfer matrix meets an output condition; If the diffraction transfer matrix meets the output condition, determining the target form to be output; Based on the target form, determining a corresponding signal feeding technology; Based on the signal feeding technology and the diffraction transmission matrix, a target form signal is output.

6. The method according to claim 5, characterized in that The method further comprises: If the diffraction transfer matrix does not meet the output condition, adjusting the parameters of the modulation kernel based on the diffraction transfer matrix; Based on the diffraction transfer matrix, the above steps are repeated until the diffraction transfer matrix meets the output conditions and the target form signal is output.

7. A reconfigurable diffraction universal intelligent optical computing system, characterized in that: The system comprises: A signal feeding module is used to obtain a first matrix corresponding to an input mode and obtain a signal parameter matrix through a modulator array; A parameter modulation module, used for obtaining a second matrix of input parameters, and processing the second matrix through a modulation kernel to obtain a corresponding modulation parameter matrix; A channel synthesis module, used for randomly fusing the modulation parameter matrix and the signal parameter matrix to obtain a fusion matrix; A diffraction propagation module, used for inputting the fusion matrix into a target diffraction propagation model to obtain a first diffraction transmission matrix; The result output module is used to output a target form signal based on the first diffraction transmission matrix.

8. The system according to claim 7, characterized in that The parameter modulation module is specifically used for: Determine the parameter type corresponding to the input parameter; Based on the parameter type, determining a target modulation parameter corresponding to the modulation core; Based on the target modulation parameters, the second matrix is ​​processed by a modulation kernel to obtain a corresponding modulation parameter matrix.

9. The system according to claim 7, characterized in that The channel synthesis module is specifically used for: The modulation parameter matrix and the signal parameter matrix are fused in parallel or cross-fused to obtain a fused matrix.

10. A reconfigurable diffraction universal intelligent optical computing architecture, characterized in that: include: At least one reconfigurable diffractive universal intelligent optical computing system as claimed in any one of claims 7 to 9.

Images