Computational Method Applied to Optical Neural Network Reconstruction-Suppression Neuron Architecture

Abstract

The present disclosure relates to the field of optical computing technologies, and particularly to a computing method applied to an optical neural network reconstruction-suppression neuron architecture. Among them, the model includes: an optical reconstruction module, configured to receive an optical input signal and perform reconfigurable modulation on the optical input signal to obtain a modulated optical signal, wherein the modulation parameters corresponding to the reconfigurable modulation are determined by the optical computing task corresponding to the optical input signal; a suppression neuron module, configured to construct a filtering channel corresponding to the optical computing task and filter the modulated optical signal according to the filtering channel to obtain and output an optical output signal. The present disclosure adopting the above solution can achieve high-performance optical computing.

Description

Technical Field

[0001] The present disclosure relates to the field of optical computing technologies, and particularly to a computing method applied to an optical neural network reconstruction-suppression neuron architecture. Background Art

[0002] With the rapid development of the fields of artificial intelligence and scientific computing, the complexity and scale of computing requirements are also increasing continuously. However, existing electronic computing technologies are limited by Moore's law, and their performance is gradually approaching the saturation state, making it difficult to effectively meet the increasingly stringent requirements of large-scale complex algorithms for computing power and power consumption. Light has natural advantages such as high throughput and low latency during propagation. Optical computing technologies that use photons instead of electrons as computing carriers are regarded as the key to breaking the existing computing bottleneck. Summary of the Invention

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

[0004] To this end, the first object of the present disclosure is to propose an optical neural network reconstruction-suppression neuron model to achieve high-performance optical computing.

[0005] The second object of the present disclosure is to propose an optical neural network reconstruction-suppression neuron architecture.

[0006] To achieve the above object, an embodiment of the first aspect of the present disclosure proposes an optical neural network reconstruction-suppression neuron model, including:

[0007] An optical reconstruction module, configured to receive an optical input signal and perform reconfigurable modulation on the optical input signal to obtain a modulated optical signal, wherein the modulation parameters corresponding to the reconfigurable modulation are determined by the optical computing task corresponding to the optical input signal;

[0008] A suppression neuron module, configured to construct a filtering channel corresponding to the optical computing task and filter the modulated optical signal according to the filtering channel to obtain and output an optical output signal.

[0009] Optionally, before the optical reconstruction module is configured to perform reconfigurable modulation on the optical input signal, it is further configured to:

[0010] Load the pre-trained weights corresponding to the optical computing task and determine the phase modulation parameters corresponding to the pre-trained weights;

[0011] Perform phase modulation on the optical input signal according to the phase modulation parameters.

[0012] Optionally, when the optical reconstruction module is configured to perform phase modulation on the optical input signal according to the phase modulation parameters, it is specifically configured to:

[0013] Determine the thermal tuning voltage corresponding to the phase modulation parameter;

[0014] Apply the thermal tuning voltage to the thermal phase modulator to control the thermal phase modulator to perform phase modulation on the optical input signal.

[0015] Optionally, when the inhibitory neuron module is used to construct the filtering channel corresponding to the optical computing task, it is specifically used for:

[0016] Input broadband light in the wavelength range corresponding to the optical computing task into an optical cavity with a tunable periodically arranged structure to obtain the first total coupled energy corresponding to the wavelength range and the second total coupled energy corresponding to outside the wavelength range;

[0017] Taking the minimum difference between the first total coupled energy and the second total coupled energy as the target, tune the optical cavity until a target optical cavity that meets the iteration requirements is obtained, so as to obtain the filtering channel corresponding to the optical computing task.

[0018] Optionally, when the inhibitory neuron module is used to filter the modulated optical signal according to the filtering channel, it is specifically used for:

[0019] Input the modulated optical signal into the target optical cavity to control the target optical cavity to filter the modulated optical signal.

[0020] Optionally, the model further includes:

[0021] A feedback control module, which is used to detect the optical output signal and optimize the modulation parameter and the filtering channel according to the detection result.

[0022] Optionally, when the feedback control module is used to detect the optical output signal and optimize the modulation parameter and the filtering channel according to the detection result, it is specifically used for:

[0023] Obtain the small distortion signal of the optical output signal caused by the environment;

[0024] Perform parameter search on the reconstruction parameter based on the distortion small signal to obtain the target reconstruction parameter;

[0025] Optimize the modulation parameter and the filtering channel according to the target reconstruction parameter.

[0026] Optionally, the model further includes:

[0027] A mode adjustment module, configured to sense the execution information of the optical computing task, and adjust the working modes of the optical reconstruction module and the inhibitory neuron module according to the execution information.

[0028] To achieve the above object, an embodiment of the second aspect of the present disclosure provides an optical neural network reconstruction-inhibitory neuron architecture, including: at least one optical neural network reconstruction-inhibitory neuron model shown in any one of the foregoing first aspects.

[0029] In summary, the optical neural network reconstruction-inhibitory neuron model and architecture provided by the present disclosure deeply integrate optical computing with the brain-like neuron method, innovatively combine and implement "reconfigurable modulation" and "reconfigurable filtering", not only improving the computing density and efficiency, but also enhancing the adaptability of the model to complex tasks, and achieving high-performance optical computing.

[0030] Additional aspects and advantages of the present disclosure will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present disclosure. Description of the Drawings

[0031] The above and / or additional aspects and advantages of the present disclosure will become apparent and understandable from the following description of the embodiments in conjunction with the drawings, where:

[0032] Figure 1 It is a schematic structural diagram of an optical neural network reconstruction-inhibitory neuron model provided by an embodiment of the present disclosure. Detailed Description of the Embodiments

[0033] The embodiments of the present disclosure will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present disclosure and should not be construed as limiting the present disclosure.

[0034] The efficient information processing ability of the brain has always been an important source of inspiration for artificial intelligence and neural network research. Especially in complex tasks such as perception, memory, and reasoning, the brain exhibits the characteristics of low power consumption, high speed, and high robustness. In recent years, neuroscience research has found that the reconstruction-inhibitory neurons in the brain play an important role in information processing and energy optimization. These neurons regulate the activities of specific neurons, ensuring both the effective expression of information and the suppression of the propagation of redundant information, thereby improving the overall efficiency of the neural network.

[0035] Although traditional electronic neural networks have achieved great success in many applications, their limitations in energy consumption and speed remain problems that need to be urgently solved. In scenarios that require real-time processing of massive amounts of data, such as autonomous driving, drone control, and biomedical signal processing, the power consumption problem of electronic neural networks is particularly prominent. In addition, as the complexity of deep learning models continues to increase, the performance bottlenecks of hardware systems have become more apparent.

[0036] As an emerging computing framework, the Optical Neural Network (ONN) shows great potential. Thanks to the high-speed propagation and high-bandwidth characteristics of light, the optical neural network can achieve parallel computing with extremely low power consumption and far exceed traditional electronic neural networks in terms of speed. However, existing optical neural network architectures often lack in-depth simulation of the mechanisms of biological neural networks, especially in terms of information selectivity and redundancy suppression. To further improve the performance of optical neural networks, introducing the mechanism of reconstruction-suppression neurons in the brain into optical neural networks and designing an optical computing architecture with adaptive information regulation ability is a highly innovative and challenging direction.

[0037] The following will describe the present disclosure in detail with specific embodiments.

[0038] It should be noted that the human brain can achieve reconfiguration by adjusting attention, separating and integrating networks, etc., to handle multiple tasks. Brain-inspired neural networks are artificial neural networks inspired by the structure and function of the human brain. They simulate certain characteristics of the biological nervous system in order to process information in a more natural and efficient way. Networks usually try to imitate the brain's learning mechanism, neuron connections, and information processing methods to improve computing power and adaptability. Intelligent optical computing networks can also achieve a brain-like structure through hardware selection and architecture design.

[0039] Figure 1 The following is a schematic structural diagram of a reconstruction-suppression neuron model of an optical neural network provided by an embodiment of the present disclosure. As Figure 1 shown, the reconstruction-suppression neuron model of the optical neural network includes:

[0040] An optical reconstruction module for receiving an optical input signal and performing reconfigurable modulation on the optical input signal to obtain a modulated optical signal;

[0041] A suppression neuron module for constructing a filtering channel corresponding to the optical computing task and filtering the modulated optical signal according to the filtering channel to obtain and output an optical output signal.

[0042] According to some embodiments, the modulation parameters corresponding to the reconfigurable modulation are determined by the optical computing task corresponding to the optical input signal, so as to reflect the reconfigurability of the neurons.

[0043] In some embodiments, the modulated optical signal is filtered according to the filtering channel corresponding to the optical computing task. Therefore, irrelevant information or redundant components in the modulated optical signal can be accurately screened out and effectively removed from the modulated optical signal, thereby improving the computational accuracy and robustness of the model. This mechanism corresponds to the role of "reconstruction-inhibition neurons" in the brain, which can dynamically adjust the information flow and ensure that the final output retains only the most relevant information.

[0044] It is easy to understand that this model deeply integrates optical computing with brain-like neuron methods, innovatively combines and implements "reconfigurable modulation" and "reconfigurable filtering" to extract the key features corresponding to the optical computing task from the input signal. This key feature can then be input as a new input signal to the next optical neural network reconstruction-inhibition neuron model. This process not only ensures the integrity of the signal and improves the computing density and efficiency, but also enhances the model's adaptability to complex tasks and achieves high-performance optical computing.

[0045] Optionally, before the optical reconstruction module is used to perform reconfigurable modulation on the optical input signal, it is also used to:

[0046] Loading the pre-trained weights corresponding to the optical computing task, and determining the phase modulation parameters corresponding to the pre-trained weights;

[0047] The optical input signal is phase modulated according to the phase modulation parameter.

[0048] According to some embodiments, the pre-trained weights are obtained by training a preset model according to an optical computing task.

[0049] In some embodiments, the phase modulation of the optical input signal can be achieved by designing a thermal phase modulator on the optical waveguide path. For example, a modulated voltage can be applied to the thermal phase modulator through the lead-out electrode. , thereby heating the area where the thermal phase modulator is located. The temperature changes the refractive index of the local area, thus affecting the optical path of the light on the waveguide. , the phase can be modulated by thermally adjustable voltage, realizing the phase reconfigurability of neurons. Secondly, since the responses of light with different coupling modes and light with different wavelengths after passing through the thermal phase adjuster are different, the input light mode and thermally adjustable voltage of the optical input signal can be jointly optimized to achieve richer neuron reconfigurability. Among them, Q The heat generated when heating the thermal phase modulator, U is the thermal regulation voltage, For the light path.

[0050] That is to say, when the optical reconstruction module is used to perform phase modulation on the optical input signal according to the phase modulation parameter, it can be used for:

[0051] Determine the thermal tuning voltage corresponding to the phase modulation parameter;

[0052] Apply the thermal tuning voltage to the thermal phase modulator to control the thermal phase modulator to perform phase modulation on the optical input signal.

[0053] Optionally, when the inhibitory neuron module is used to construct the filtering channel corresponding to the optical computing task, it is specifically used for:

[0054] Input the broadband light in the wavelength range corresponding to the optical computing task into the optical cavity with a tunable periodically arranged structure to obtain the first total coupled energy corresponding to the wavelength range and the second total coupled energy corresponding to outside the wavelength range;

[0055] Taking the minimum difference between the first total coupled energy and the second total coupled energy as the goal, tune the optical cavity until a target optical cavity that meets the iteration requirements is obtained, so as to obtain the filtering channel corresponding to the optical computing task.

[0056] Taking a scenario as an example, when the wavelength range corresponding to the optical computing task is from 1.5 microns to 1.6 microns, the difference between the first total coupled energy and the second total coupled energy E Can be determined according to the following formula:

[0057]

[0058] According to some embodiments, during the process of tuning the optical cavity, the particle swarm genetic algorithm can be used for optimization iteration to solve the minimum value of the difference to search for the optimal design of the optical cavity, so as to obtain the target optical cavity.

[0059] In some embodiments, the inhibitory neuron module can input the modulated optical signal into the target optical cavity to control the target optical cavity to filter the modulated optical signal.

[0060] It should be noted that the initial optical input signal can be preprocessed and compressed into a one-dimensional signal, coupled through the tunable periodically arranged structure optical cavity of the neuron, and the on-chip light intensity absorption curve can be selectively regulated, so as to realize the preliminary encoding of optical information and obtain the optical input signal.

[0061] It is easy to understand that the inhibitory neuron module realizes reconfigurable filtering by adopting the functional hardware design based on photonic crystals, and can realize richer neuron reconfigurability.

[0062] Optionally, the optical neural network reconstruction - inhibitory neuron model further includes:

[0063] A feedback control module for detecting the optical output signal and optimizing the modulation parameters and filtering channels according to the detection results.

[0064] It should be noted that by introducing an adaptive adjustment mechanism based on feedback control, the characteristics of the optical input signal are detected in real time based on the detection of the optical output signal, and the parameter configurations of the optical reconstruction module and the inhibitory neuron module are automatically adjusted according to the change of characteristics, so as to realize the dynamic adjustment of the optical neural network reconstruction-inhibitory neuron model, enabling the model to flexibly respond to different task requirements, maintain excellent performance, and ensure that the entire model can operate efficiently in a changing task environment.

[0065] According to some embodiments, when the feedback control module is used to detect the optical output signal and optimize the modulation parameters and filtering channels according to the detection results, it is specifically used for:

[0066] Obtain the distorted small signal of the optical output signal caused by the environment;

[0067] Perform parameter search on the reconstruction parameters based on the distorted small signal to obtain the target reconstruction parameters;

[0068] Optimize the modulation parameters and filtering channels according to the target reconstruction parameters.

[0069] In some embodiments, a parameter search based on the genetic algorithm can be performed on the reconstruction parameters based on a preset maximum search step, and this preset maximum search step does not specifically refer to a certain fixed step. For example, this preset maximum search step can be 10.

[0070] In some embodiments, the modulation parameters and filtering channels can be optimized by optimizing the weight distribution of the optical modulation and filtering processes according to the target reconstruction parameters.

[0071] Optionally, the optical neural network reconstruction-inhibitory neuron model further includes:

[0072] A mode adjustment module for perceiving the execution information of the optical computing task and adjusting the working modes of the optical reconstruction module and the inhibitory neuron module according to the execution information.

[0073] According to some embodiments, the execution information includes but is not limited to information such as computing accuracy and running duration.

[0074] Taking a scenario as an example, in an image processing task, the mode adjustment module can dynamically optimize the interaction modes of the optical reconstruction module and the inhibitory neuron module according to different scenarios. For example, if it senses that the optical scene coupled into the network changes little, the maximum search step of the search algorithm can be reduced, so as to obtain the best processing speed and the lowest power consumption.

[0075] It should be noted that by means of "reinforcement learning", the execution situation of the optical computing task is sensed, and the entire optical input signal processing path is intelligently adjusted, that is, the working modes of the optical reconstruction module and the inhibitory neuron module are automatically adjusted. Therefore, the robustness and adaptability of the model can be enhanced.

[0076] In summary, the model provided in this embodiment is inspired by the fully reconfigurable mechanism of the human brain, realizes "reconfigurable modulation" and "reconfigurable filtering" of the input information, thereby significantly improving the computing density and accuracy while maintaining high-speed computing. At the same time, it can achieve adaptive reconfigurable adjustment, improve the robustness and efficiency of network inference, and effectively reduce the system energy consumption. This innovative optical architecture has broad application prospects, such as in edge computing, real-time signal processing, and artificial intelligence applications in energy-constrained devices.

[0077] To implement the above embodiment, the present disclosure also proposes an optical neural network reconstruction-inhibitory neuron architecture, including: at least one optical neural network reconstruction-inhibitory neuron model provided in the foregoing embodiment.

[0078] In summary, the architecture provided in this embodiment not only has high-efficiency parallel computing capabilities, but also can maintain stable performance in complex tasks and changing environments, which can lay a solid foundation for the development of the next generation of artificial intelligence hardware.

[0079] The collection, storage, use, processing, transmission, provision, and disclosure of the user's personal information involved in the present disclosure and other processing are all in compliance with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0080] It should be noted that personal information from users should be collected for legal and reasonable purposes and should not be shared or sold outside of these legal uses. In addition, such collection / sharing should be carried out after obtaining the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and signing an agreement / authorization including authorizing relevant user information before the user uses the function. In addition, any necessary steps need to be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0081] The present disclosure anticipates providing embodiments for users to selectively block the use or access of personal information data. That is, the present disclosure anticipates providing hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, the risk can be minimized by restricting data collection and deleting data. In addition, when applicable, personal identifiers are removed from such personal information to protect the privacy of users.

[0082] In the technical solution of the present disclosure, the acquisition, transmission, storage, use, processing, etc. of data all comply with the relevant provisions of national laws and regulations.

[0083] It should be noted that in the embodiments of the present disclosure, some existing solutions in the industry such as certain software, components, models, etc. may be mentioned. They should be regarded as exemplary. The purpose is only to illustrate the feasibility in the implementation of the technical solution of this application, but it does not mean that the applicant has already or necessarily used this solution.

[0084] In the description of the foregoing embodiments, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations 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, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0085] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0086] Any process or method description shown in the flowchart or described in other ways herein can be understood to represent a module, segment, or part of code including one or more executable instructions for implementing a customized logical function or process. The scope of the preferred embodiments of the present disclosure includes additional implementations, where the functions may be executed in a manner that is not shown or discussed, including in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by those skilled in the art to which the embodiments of the present disclosure belong.

[0087] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definable sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which a program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then storing it in a computer memory.

[0088] It should be understood that various parts of the present disclosure can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having suitable combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0089] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by a program instructing relevant hardware. 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 embodiments.

[0090] In addition, each functional unit in various embodiments of the present disclosure may be integrated into a processing module, may exist physically alone for each unit, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0091] The above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A computational method for reconstructing and inhibiting neuron architecture in an optical neural network, characterized in that: The optical neural network reconstruction-inhibition neuron architecture includes at least one optical neural network reconstruction-inhibition neuron model, the optical neural network reconstruction-inhibition neuron model includes an optical reconstruction module and an inhibitory neuron module, and the method includes: Controlling the optical reconstruction module to receive an optical input signal and to perform reconfigurable modulation on the optical input signal to obtain a modulated optical signal, wherein a modulation parameter corresponding to the reconfigurable modulation is determined by an optical computing task corresponding to the optical input signal; Controlling the inhibitory neuron module to construct a filter channel corresponding to the optical computing task, and filtering the modulated optical signal according to the filter channel to obtain and output an optical output signal; Before controlling the optical reconstruction module to perform reconfigurable modulation on the optical input signal, a pre-trained weight corresponding to the optical computing task is also loaded, and a phase modulation parameter corresponding to the pre-trained weight is determined to perform phase modulation on the optical input signal according to the phase modulation parameter; When controlling the optical reconstruction module to phase modulate the optical input signal according to the phase modulation parameter, determining a thermal adjustment voltage corresponding to the phase modulation parameter; applying the thermal adjustment voltage to the thermal phase modulator to control the thermal phase modulator to phase modulate the optical input signal; When controlling the inhibitory neuron module to construct the filtering channel corresponding to the optical computing task, broadband light in the wavelength range corresponding to the optical computing task is input into an optical cavity of a tunable periodic arrangement structure to obtain a first coupling total energy corresponding to the wavelength range and a second coupling total energy corresponding to outside the wavelength range; with the goal of minimizing the difference between the first coupling total energy and the second coupling total energy, the optical cavity is tuned until a target optical cavity that meets the iteration requirements is obtained, so as to obtain the filtering channel corresponding to the optical computing task.

2. The method according to claim 1, characterized in that: The filtering of the modulated optical signal according to the filtering channel comprises: The modulated optical signal is input into the target optical cavity to control the target optical cavity to filter the modulated optical signal.

3. The method according to claim 1, characterized in that The optical neural network reconstruction-inhibition neuron model also includes a feedback control module, and the method also includes: The feedback control module is controlled to detect the optical output signal, and the modulation parameter and the filter channel are optimized according to the detection result.

4. The method according to claim 3, characterized in that The controlling the feedback control module to detect the optical output signal and optimizing the modulation parameter and the filter channel according to the detection result includes: Obtaining a small signal of distortion caused by the optical output signal to the environment; Performing parameter search on reconstruction parameters based on the distorted small signal to obtain target reconstruction parameters; The modulation parameters and the filter channel are optimized according to the target reconstruction parameters.

5. The method according to claim 1, characterized in that The optical neural network reconstruction-inhibition neuron model also includes a mode adjustment module, and the method also includes: The mode adjustment module is controlled to sense the execution information of the optical computing task, and the working modes of the optical reconstruction module and the neuron inhibition module are adjusted according to the execution information.

Images