Completely reconfigurable universal intelligent optical computing chip architecture and system

Abstract

The invention relates to the technical field of optical computing, in particular to a fully reconfigurable universal intelligent optical computing chip architecture and system. The framework comprises a transmitting module used for performing beam splitting on each kind of first single-wavelength light in first broadband light corresponding to a target task to obtain and transmit at least one single-wavelength beam splitting light set; the receiving module is used for receiving the at least one single-wavelength beam splitting light set and carrying out matrix calculation on the at least one single-wavelength beam splitting light set according to the optical propagation matrix corresponding to the target task to obtain second broadband light after light calculation, the single-wavelength beam splitting light sets are in one-to-one correspondence with rows in the optical propagation matrix, and the rows in the optical propagation matrix are in one-to-one correspondence with the rows in the target task. And the single-wavelength split light in the single-wavelength split light set is in one-to-one correspondence with the row elements in the rows. According to the invention, the method can achieve the integrated and reconfigurable operation of a large-scale optical matrix through a single layer at the same time.

Description

Technical Field

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

[0002] With the rapid development of artificial intelligence and scientific computing, the complexity and scale of computing needs are also increasing. However, the existing electronic computing technology is limited by Moore's Law, and its performance is gradually approaching saturation, making it difficult to effectively cope with the increasingly stringent requirements of large-scale complex algorithms on computing power and power consumption. Light has natural advantages such as high throughput and low latency in the propagation process. Optical computing technology that uses photons instead of electrons as computing carriers is seen as the key to breaking the existing computing bottleneck. Summary of the invention

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

[0004] To this end, the first objective of the present disclosure is to propose a fully reconfigurable universal intelligent optical computing chip architecture to achieve integrated and reconfigurable large-scale optical matrix operations using a single layer.

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

[0006] To achieve the above objectives, the first embodiment of the present disclosure proposes a fully reconfigurable universal intelligent optical computing chip architecture, including: A transmitting module, used for splitting each first single wavelength light in the first broadband light corresponding to the target task, obtaining and transmitting at least one single wavelength split light set, wherein the first single wavelength light corresponds to the single wavelength split light set one by one; A receiving module is used to receive the at least one single-wavelength split light set, and perform matrix calculation on the at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task to obtain a second broadband light after optical calculation, wherein the second broadband light includes at least one second single-wavelength light, the second single-wavelength light corresponds to the first single-wavelength light in a one-to-one manner, the single-wavelength split light set corresponds to a row in the optical propagation matrix in a one-to-one manner, and the single-wavelength split lights in the single-wavelength split light set correspond to row elements in the row in a one-to-one manner.

[0007] Optionally, the transmitting module is used to split each first single-wavelength light in the first broadband light corresponding to the target task to obtain at least one single-wavelength split light set, specifically for: Splitting each first single-wavelength light in the first broadband light corresponding to the target task to obtain at least one initial single-wavelength split light set; The at least one initial single-wavelength split light set is modulated according to the target task to obtain at least one single-wavelength split light set.

[0008] Optionally, when the transmitting module is used to modulate the at least one initial single-wavelength split light set according to the target task, it is specifically used to: The at least one initial single-wavelength split light beam set is phase modulated according to the target task.

[0009] Optionally, before performing matrix calculation on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, the receiving module is further used to: Determine the number of wavelengths corresponding to the first single wavelength light in the target task, and the number of split beams corresponding to the single wavelength split beam set; Constructing an initial optical propagation matrix using the number of wavelengths as the number of rows and the number of split beams as the number of columns; According to the target task, the value of each element in the initial optical propagation matrix is ​​determined to obtain the optical propagation matrix corresponding to the target task.

[0010] Optionally, when the receiving module is used to determine the value of each element in the initial optical propagation matrix according to the target task, it is specifically used to: According to the target task, the value of each element in the initial optical propagation matrix is ​​trained.

[0011] Optionally, when the receiving module is used to perform matrix calculation on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, it is specifically used to: The light intensity corresponding to each single-wavelength split light beam in the at least one single-wavelength split light beam set is used as input to perform a matrix multiplication operation with the optical propagation matrix.

[0012] Optionally, when the transmitting module is used to transmit the at least one single-wavelength split light set, it is specifically used to: The at least one single wavelength split light set is emitted through an emission grating.

[0013] To achieve the above-mentioned purpose, the second aspect of the present disclosure proposes a fully reconfigurable universal intelligent optical computing chip system, including: at least one fully reconfigurable universal intelligent optical computing chip architecture shown in any one of the aforementioned first aspects.

[0014] Optionally, the architecture includes: a plurality of the fully reconfigurable universal intelligent optical computing chip architectures.

[0015] Optionally, the connection mode between the multiple fully reconfigurable universal intelligent optical computing chip architectures includes at least one of the following: Series connection; in parallel.

[0016] In summary, the fully reconfigurable general intelligent optical computing chip architecture and system provided by the present invention, when performing matrix calculations on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, optical calculations are performed by using the single-wavelength split light in the single-wavelength split light set and the corresponding row elements, thereby breaking through the contradiction between reconfigurability and large-scale integration, and achieving integrated and reconfigurable large-scale optical matrix operations using a single layer.

[0017] 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 learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A schematic diagram of the structure of a fully reconfigurable universal intelligent optical computing chip architecture provided by an embodiment of the present disclosure; Figure 2 A schematic diagram of the principle of a fully reconfigurable universal intelligent optical computing chip architecture provided by an embodiment of the present disclosure; Figure 3 A simulation result diagram of a fully reconfigurable universal intelligent optical computing chip architecture provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] Embodiments of the present disclosure are described in detail below, and examples of the embodiments 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.

[0020] With the rapid development of artificial intelligence technology, especially the breakthroughs in large models (such as large language models, cultural video models, etc.), computing needs are growing exponentially. The scale of large models has gradually grown from millions of parameters in the early days to tens of billions or even trillions of parameters today. These models have shown great potential in natural language processing, computer vision, autonomous driving and other fields. However, training and reasoning such large-scale models pose unprecedented challenges to computing power. The bottleneck of traditional electronic computing architecture is its high power consumption and limited computing speed, which cannot fully meet the growing demand for large-scale computing of artificial intelligence models.

[0021] In addition, the current optical computing architecture faces the contradiction of computing scale and reconfigurability. At the same time, the optical chips that have been designed have fixed computing modes and cannot have the ability to reconfigure computing modes. In addition, optical chip computing cannot realize arbitrary matrices in the case of single-layer propagation, and usually requires multi-layer multiplexing to complete. These limitations make optical computing unable to meet the current large-model computing needs that require reconfiguration and large-scale integration, and fail to fully realize the potential of optical computing.

[0022] Therefore, how to make optical computing chips both integrated and reconfigurable and use a single layer to complete large-scale matrix operations has become a technical problem that needs to be solved urgently.

[0023] The present disclosure is described in detail below with reference to specific embodiments.

[0024] Figure 1 This is a schematic diagram of a fully reconfigurable universal intelligent optical computing chip architecture provided by an embodiment of the present disclosure. Figure 1 As shown, the fully reconfigurable universal intelligent optical computing chip architecture includes: A transmitting module, used for splitting each first single-wavelength light in the first broadband light corresponding to the target task, to obtain and transmit at least one single-wavelength split light set; The receiving module is used to receive at least one single-wavelength split light set, and perform matrix calculation on the at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task to obtain the second broadband light after the optical calculation.

[0025] According to some embodiments, a high-speed signal corresponding to a target task may be loaded into a broadband spectrum to obtain a first broadband light, and the first broadband light may be multiplexed into a transmitting module.

[0026] In some embodiments, the first single wavelength light corresponds to the single wavelength split light set one by one; in the transmitting module, each first single wavelength light in the input first broadband light needs to be transmitted on the same waveguide and emitted for diffraction, so as to be split from one waveguide to multiple waveguides, and obtain a single wavelength split light set composed of multiple single wavelength split lights. In the process of splitting, even division may be performed or not, which can be adjusted according to the application scenario.

[0027] According to some embodiments, the second broadband light includes at least one second single wavelength light, and the second single wavelength light corresponds to the first single wavelength light one by one. The "first" in the first broadband light, the second broadband light, the first single wavelength light, and the second single wavelength light, and the "second" in the second integrated transmission and computing conversion chip have no special meanings and are only used for distinction.

[0028] In some embodiments, the single-wavelength split light set corresponds to a row in the optical propagation matrix one-to-one, and the single-wavelength split light in the single-wavelength split light set corresponds to a row element in the row one-to-one. That is, when performing matrix calculation on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, the single-wavelength split light in the single-wavelength split light set performs optical calculation with the corresponding row element.

[0029] It is easy to understand that when this architecture performs matrix calculations on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, optical calculations are performed between the single-wavelength split light in the single-wavelength split light set and the corresponding row elements, thereby breaking through the contradiction between reconfigurability and large-scale integration, and achieving integration and reconfiguration while completing large-scale optical matrix operations using a single layer.

[0030] Optionally, the transmitting module is used to split each first single-wavelength light in the first broadband light corresponding to the target task to obtain at least one single-wavelength split light set, specifically for: Splitting each first single-wavelength light in the first broadband light corresponding to the target task to obtain at least one initial single-wavelength split light set; At least one initial single-wavelength split light set is modulated according to a target task to obtain at least one single-wavelength split light set.

[0031] According to some embodiments, when the transmitting module is used to modulate at least one initial single-wavelength split light set according to the target task, the at least one initial single-wavelength split light set may be phase modulated according to the target task.

[0032] In some embodiments, an Application Specific Integrated Circuit (ASIC) interface may be used to perform phase modulation on at least one initial single-wavelength split light beam set.

[0033] In some embodiments, the voltage source may be driven to drive the ASIC interface through the electrode to achieve phase modulation of the initial single-wavelength split light set.

[0034] Optionally, when the transmitting module is used to transmit at least one single-wavelength split light set, it is specifically used to: At least one single wavelength split light set is emitted through the emission grating.

[0035] According to some embodiments, the single-wavelength split light set may be emitted from the grating end of the corresponding transmitting grating and received by a receiving module of the next stage.

[0036] Optionally, before performing matrix calculation on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, the receiving module is further used to: Determine the number of wavelengths corresponding to the first single wavelength light in the target task, and the number of split beams corresponding to the single wavelength split beam light set; The initial optical propagation matrix is ​​constructed with the number of wavelengths as the number of rows and the number of beam splitting as the number of columns; According to the target task, the value of each element in the initial optical propagation matrix is ​​determined to obtain the optical propagation matrix corresponding to the target task.

[0037] It should be noted that Figure 2 The schematic diagram of the principle of a fully reconfigurable universal intelligent optical computing chip architecture provided by the embodiment of the present disclosure. Figure 2 As shown in the figure, the element values ​​of the traditional optical computing matrix architecture are very close when the wavelength interval is close, so the column vectors of the matrix are highly linearly correlated, which makes the matrix rank of the chip implementation small. The multi-wavelength decoupling theory is derived from the decoupling of different wavelengths of the shared waveguide on the chip waveguide and the spatial phase decoupling of waveguides at different spatial positions. For a single transmitting waveguide, the distance propagated on the waveguide from the input light to the final emission is , this distance will make the phase decoupling of light of different wavelengths at the time of emission random, because after the wavelength propagation and The phase difference between the two wavelengths can be described as The long waveguide causes the original close phase difference to be amplified, so that the initial phase of different wavelengths emitted by a single waveguide is inconsistent; for the outgoing light at different positions, after the previous emission grating, the waveguide light is modulated by the waveguide for a period of time to respond to different wavelengths with different phases. , thereby realizing the outgoing light on the entire emission surface (assuming there are 256 emission gratings in total). For light of different wavelengths, the corresponding propagation matrices are highly nonlinearly correlated, thus providing theoretical and architectural support for the single-layer realization of arbitrary matrices.

[0038] In some embodiments, Figure 2 As shown in the figure, existing optical computing only reconstructs trainable parameters, and its architecture is fixed and cannot be reconstructed. At the same time, the computing scale can only be achieved under limited chip area. The architecture provided in this embodiment can realize matrix input under the condition of obtaining effective waveguide multi-wavelength phase perturbation. Output In meeting the conditions In the case of full freedom, the present invention can DoF The input dimension depends on the number of input wavelengths, and the wavelength interval should meet the minimum perturbation standard. , and the output dimension can be adaptively adjusted by changing the number of receiving positions. That is to say, the architecture provided by this embodiment can realize general intelligent optical computing with reconfigurable matrix architecture, including reconfigurable parameters. For example, Arbitrary matrix operations with full degrees of freedom.

[0039] In some embodiments, Figure 3 This is a simulation result diagram of a fully reconfigurable universal intelligent optical computing chip architecture provided by an embodiment of the present disclosure. Figure 3 As shown in , it shows that the relative error of the realized matrix is ​​significantly reduced in the case of chip phase disturbance compared to the case without disturbance. Through simulation and experiment, it is proved that the input and output degrees of freedom of any matrix and the realization accuracy of the realization matrix are as follows Figure 3 The relationship shown, in which the dividing line function between high and low precision is , the parts below this function are all implemented with effective high-precision matrices.

[0040] According to some embodiments, when the receiving module is used to determine the value of each element in the initial optical propagation matrix according to the target task, the value of each element in the initial optical propagation matrix may be trained according to the target task.

[0041] In some embodiments, the value of the element is the response value of the corresponding single-wavelength split light at unit intensity. The value can also be determined according to the following formula:

[0042] Among them, the single wavelength split light set includes N A single wavelength split beam, N is a positive integer greater than 1, W i is the first i The value of the element corresponding to a single wavelength split beam. G 0i Refers to the first i The propagation matrix of a single-wavelength split light beam in space. Refers to the first i The optical propagation parameters corresponding to a single wavelength split light.

[0043] It is easy to understand that by designing waveguide spacing spectral perturbations on the chip and jointly modeling the space and spectrum of light at the input end, a fully reconfigurable optical computing chip spatial arbitrary-dimensional diffraction calculation model can be constructed, which can realize a single-layer optical computing chip with a matrix of arbitrary shape with full degrees of freedom, breaking the contradiction between reconfigurability and large-scale integration.

[0044] Optionally, when the receiving module is used to perform matrix calculation on at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, it is specifically used to: The light intensity corresponding to each single-wavelength split light beam in at least one single-wavelength split light beam set is used as input to perform a matrix multiplication operation with the optical propagation matrix.

[0045] Taking a scenario as an example, when the first broadband light includes only one first single-wavelength light and the first single-wavelength light is split into N single-wavelength split lights, the input of the optical propagation matrix can be expressed as follows:

[0046] in, x Indicates the intensity of single wavelength split light.

[0047] The optical propagation matrix corresponding to the first single wavelength light can be expressed as follows:

[0048] Next, X and W The second single wavelength light obtained after matrix multiplication can be expressed as follows:

[0049] According to some embodiments, a matrix multiplication operation is performed to obtain x 1 W 1. x 2 W 2 …x i W i Finally, they can be added together through incoherent superposition to complete the calculation process of the entire matrix.

[0050] In summary, the architecture provided in this embodiment is supported by the multi-wavelength decoupling theory, and uses the wavelength information of the light field to achieve phase decoupling through waveguide perturbations on the chip, thereby realizing spatial phase decoupling, thereby maximizing the use of modulation degrees of freedom, and achieving fully reconfigurable general matrix computing that reaches the upper limit of theoretical matrix computing capabilities. It can greatly expand the application scope of photonic computing, lay a solid foundation for optical computing of large artificial intelligence models, and is expected to bring new opportunities for high-performance artificial intelligence large model computing in the post-Moore era, and realize the optical chip-side deployment of large artificial intelligence models.

[0051] In order to implement the above embodiments, the present disclosure also proposes a fully reconfigurable universal intelligent optical computing chip system, including: a fully reconfigurable universal intelligent optical computing chip architecture provided by at least one of the above embodiments.

[0052] Optionally, the architecture includes: a plurality of fully reconfigurable universal intelligent optical computing chip architectures. The connection mode between the plurality of fully reconfigurable universal intelligent optical computing chip architectures includes at least one of the following: Series connection; in parallel.

[0053] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this disclosure shall comply with the relevant laws and regulations and shall not violate public order and good morals.

[0054] It should be noted that personal information from users should be collected for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. In addition, such collection / sharing should be carried out after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign the agreement / authorization including authorization of relevant user information before the user uses the function. In addition, any necessary steps should be taken to protect and safeguard access to such personal information data and ensure that others who have access to personal information data comply with its privacy policy and procedures.

[0055] The present disclosure anticipates providing implementation schemes 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 the personal information data is no longer needed, risks can be minimized by limiting data collection and deleting the data. In addition, when applicable, such personal information is de-identified to protect the privacy of the user.

[0056] The acquisition, transmission, storage, use, and processing of data in the technical solution disclosed in this disclosure are in compliance with the relevant provisions of national laws and regulations.

[0057] It should be noted that in the embodiments of the present disclosure, certain software, components, models and other existing solutions in the industry may be mentioned, which should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0058] In the description of the aforementioned embodiments, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means 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 representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they contradict each other.

[0059] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0060] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes 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 not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present disclosure belong.

[0061] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, 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 in a suitable manner if necessary, and then stored in a computer memory.

[0062] It should be understood that the various parts of the present disclosure can be implemented in hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented in 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, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0063] A person skilled in the art may understand that all or part of the steps in the above-mentioned embodiment method may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0064] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, 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. If 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.

[0065] The storage medium mentioned above may be a read-only memory, a disk or an optical disk, 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 cannot be understood as limitations of the present disclosure. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present disclosure.

Claims

1. A fully reconfigurable general-purpose intelligent optical computing chip architecture, characterized in that: include: A transmitting module, used for splitting each first single wavelength light in the first broadband light corresponding to the target task, obtaining and transmitting at least one single wavelength split light set, wherein the first single wavelength light corresponds to the single wavelength split light set one by one; A receiving module is used to receive the at least one single-wavelength split light set, and perform matrix calculation on the at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task to obtain a second broadband light after optical calculation, wherein the second broadband light includes at least one second single-wavelength light, the second single-wavelength light corresponds to the first single-wavelength light in a one-to-one manner, the single-wavelength split light set corresponds to a row in the optical propagation matrix in a one-to-one manner, and the single-wavelength split lights in the single-wavelength split light set correspond to row elements in the row in a one-to-one manner.

2. The architecture according to claim 1, characterized in that The transmitting module is used to split each first single wavelength light in the first broadband light corresponding to the target task to obtain at least one single wavelength split light set, specifically used to: Splitting each first single-wavelength light in the first broadband light corresponding to the target task to obtain at least one initial single-wavelength split light set; The at least one initial single-wavelength split light set is modulated according to the target task to obtain at least one single-wavelength split light set.

3. The architecture according to claim 2, characterized in that: When the transmitting module is used to modulate the at least one initial single-wavelength split light set according to the target task, it is specifically used to: The at least one initial single-wavelength split light beam set is phase modulated according to the target task.

4. The architecture according to claim 1, characterized in that Before performing matrix calculation on the at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, the receiving module is further used to: Determine the number of wavelengths corresponding to the first single wavelength light in the target task, and the number of split beams corresponding to the single wavelength split beam set; Constructing an initial optical propagation matrix using the number of wavelengths as the number of rows and the number of split beams as the number of columns; According to the target task, the value of each element in the initial optical propagation matrix is ​​determined to obtain the optical propagation matrix corresponding to the target task.

5. The architecture according to claim 4, characterized in that When the receiving module is used to determine the value of each element in the initial optical propagation matrix according to the target task, it is specifically used to: According to the target task, the value of each element in the initial optical propagation matrix is ​​trained.

6. The architecture according to claim 1, characterized in that: When the receiving module is used to perform matrix calculation on the at least one single-wavelength split light set according to the optical propagation matrix corresponding to the target task, it is specifically used to: The light intensity corresponding to each single-wavelength split light beam in the at least one single-wavelength split light beam set is used as input to perform a matrix multiplication operation with the optical propagation matrix.

7. The architecture according to claim 1, characterized in that: When the transmitting module is used to transmit the at least one single-wavelength split light set, it is specifically used to: The at least one single wavelength split light set is emitted through an emission grating.

8. A fully reconfigurable universal intelligent optical computing chip system, characterized in that: include: At least one fully reconfigurable universal intelligent optical computing chip architecture as described in any one of claims 1 to 7.

9. The system according to claim 8, characterized in that include: Multiple fully reconfigurable general intelligent optical computing chip architectures.

10. The system according to claim 9, characterized in that The connection mode between the multiple fully reconfigurable universal intelligent optical computing chip architectures includes at least one of the following: Series connection; in parallel.

Images