Diffraction neural network and three-dimensional diffraction neural network

By using a focused light control system in diffraction neural network and three-dimensional diffraction neural network to efficiently and parallelly regulate phase change materials, the problem of low regulation efficiency in the existing technology is solved, and the training efficiency and calculation speed are significantly improved.

CN118917371APending Publication Date: 2024-11-08SHENZHEN METALENX TECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411023504.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the prior art, the method of regulating phase change materials through high-power microscope objectives is inefficient, resulting in a long training time for reconfigurable diffraction neural networks and a long time for processing multiple light calculation tasks, which seriously affects the efficiency of light calculation.

Method used

Using a diffraction neural network and a three-dimensional diffraction neural network based on phase change materials, the phase change materials in the diffraction layer are efficiently controlled in parallel through the focus light control system, and multiple reconfigurables are achieved by using a multi-focus or single-focus focusing light control system. Independent regulation of micro-nano units.

Benefits of technology

The training efficiency of diffraction neural networks and three-dimensional diffraction neural networks and the processing speed of optical computing tasks have been greatly improved, and the computing efficiency and flexibility have been improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917371A_ABST
    Figure CN118917371A_ABST
Patent Text Reader

Abstract

The invention provides a diffraction neural network and a three-dimensional diffraction neural network. The diffraction neural network provided by the invention comprises an input port, a diffraction system and an output port, wherein the input port is used for receiving input light and guiding the input light to the diffraction system; the diffraction system comprises a diffraction layer and a focusing light control system; the diffraction layer comprises a metasurface, and the metasurface is used for modulating the input light; the metasurface comprises at least two reconfigurable micro-nano units based on the phase change material; the focusing light control system is used for performing phase change regulation and control on the phase change material in the diffraction layer by controlling laser; and the output port is used for receiving the output light from the diffraction layer subjected to phase change regulation and control by the focusing light control system and transmitting the output light to the detector array for receiving the output light. The diffraction neural network provided by the invention is high in training efficiency, and the speed of processing the light calculation task is accelerated.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of optical chips, and in particular to a diffraction neural network and a three-dimensional diffraction neural network. Background Art

[0002] With the development of optical technology and metamaterial research, people have carried out various studies on using optical structures to simulate neural networks, that is, optical neural networks. Among them, diffractive optical neural networks (DONNs) are a representative of optical neural networks. They use light as an information medium and use the diffraction of light to realize the calculation and processing of information, thereby performing the tasks matched during training.

[0003] In the prior art, the use of phase change materials as the diffraction layer of the reconfigurable diffractive neural network has achieved certain results. The phase change material can switch between the amorphous state and the crystalline state under the stimulation of external conditions, thereby realizing the regulation of the equivalent refractive index of the phase change material. When training the reconfigurable diffractive neural network, or when the reconfigurable diffractive neural network needs to be used to process multiple different optical computing tasks, it is necessary to regulate the multiple phase change materials in the diffraction layer of the reconfigurable diffractive neural network to meet different needs. However, in the existing method of regulating the phase change material by a high-power microscope objective lens, if you want to achieve the regulation of phase change materials at multiple different positions, you need to control the high-power microscope objective lens to translate and rotate, so that the high-power microscope objective lens can regulate each phase change material in turn to complete the regulation of all phase change materials. This regulation method takes a long time and has low regulation efficiency, which makes the training time of the reconfigurable diffractive neural network long, and the processing time of large optical computing tasks containing multiple different tasks is long, which seriously affects the efficiency of optical computing. Summary of the invention

[0004] One purpose of the present application is to propose a diffraction neural network and a three-dimensional diffraction neural network, which can realize efficient parallel control of the diffraction layer composed of phase change materials in the diffraction neural network and the three-dimensional diffraction neural network, improve the training efficiency of the diffraction neural network and the three-dimensional diffraction neural network, and improve the speed of the diffraction neural network and the three-dimensional diffraction neural network in processing optical computing tasks.

[0005] According to one aspect of an embodiment of the present application, a diffractive neural network is disclosed, the diffractive neural network comprising: an input port, a diffractive system, and an output port;

[0006] Wherein, the input port is used to receive input light and guide the input light to the diffraction system;

[0007] The diffraction system includes a diffraction layer and a focusing light control system; the diffraction layer includes a metasurface, and the metasurface is used to modulate the input light; the metasurface includes at least two reconfigurable micro-nano units based on phase change materials;

[0008] The focusing light control system is used to control the phase change of the phase change material in the diffraction layer by controlling the laser;

[0009] The output port is used to receive the output light from the diffraction layer that is phase-changed and regulated by the focusing light control system, and emit the output light to a detector array for receiving the output light.

[0010] In an exemplary embodiment of the present application, the diffraction layer includes at least two layers of metasurfaces, and the metasurfaces are arranged in sequence and spaced apart along the propagation direction of the light path.

[0011] In an exemplary embodiment of the present application, the reconfigurable micro-nano unit includes a substrate, a micro-nano structure and a spacer structure;

[0012] Wherein, the material of the micro-nano structure is a phase change material.

[0013] In an exemplary embodiment of the present application, the reconfigurable micro-nano unit includes a substrate, a micro-nano structure and a spacer structure;

[0014] Wherein, the material of the spacing structure is phase change material.

[0015] In an exemplary embodiment of the present application, the diffraction layer is disposed at one focal length of the focusing light control system.

[0016] In an exemplary embodiment of the present application, the focusing light control system is a multi-focus focusing light control system, or a single-focus focusing light control system.

[0017] In an exemplary embodiment of the present application, the multi-focal point focusing light control system includes a phase modulation module and a lens module;

[0018] The phase modulation module is used to modulate the wavefront of the control laser and transmit the control laser after wavefront modulation to the lens module;

[0019] The lens module is used to focus the control laser after wavefront modulation onto the phase change material of the diffraction layer.

[0020] In an exemplary embodiment of the present application, the phase modulation module is a spatial light modulator or a digital micromirror device;

[0021] The phase modulation module is arranged at the entrance pupil of the lens module.

[0022] According to another aspect of an embodiment of the present application, a three-dimensional diffraction neural network is disclosed, wherein the three-dimensional diffraction neural network includes a three-dimensional input port, a three-dimensional diffraction system, and a three-dimensional output port;

[0023] Wherein, the three-dimensional input port comprises at least two input ports as described in any one of the above items;

[0024] The three-dimensional diffraction system comprises at least two diffraction layers as described in any one of the above items;

[0025] The three-dimensional diffraction system further comprises at least two focusing light control systems as described in any one of the above items;

[0026] The three-dimensional output port comprises at least two output ports as described in any one of the above items.

[0027] In an exemplary embodiment of the present application, the focusing light control system and the diffraction layer are matched one by one;

[0028] The diffraction layer is arranged at one focal length of the focusing light control system.

[0029] In an exemplary embodiment of the present application, when the focal lengths of all the focusing light control systems are the same, all the focusing light control systems are not located in the same plane.

[0030] In an exemplary embodiment of the present application, when the focal lengths of all the focusing light control systems are different, all the focusing light control systems are located in the same plane.

[0031] The diffraction neural network provided in the present application includes: an input port, a diffraction system and an output port; wherein the input port is used to receive input light and guide the input light to the diffraction system; the diffraction system includes a diffraction layer and a focusing light control system; the diffraction layer includes a metasurface, and the metasurface is used to modulate the input light; the metasurface includes at least two reconfigurable micro-nano units based on phase change materials; the focusing light control system is used to control the phase change of the phase change material in the diffraction layer by controlling the laser; the output port is used to receive the output light from the diffraction layer phase-changed by the focusing light control system, and emit the output light to a detector array for receiving the output light. The diffraction neural network and three-dimensional diffraction neural network provided in the present application can independently and efficiently perform parallel phase change control on the reconfigurable micro-nano units based on phase change materials, improve the efficiency of diffraction neural network training, and speed up the processing of optical computing tasks.

[0032] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0033] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The above and other objects, features and advantages of the present application will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings.

[0035] Figure 1 A first schematic diagram of a diffractive neural network in an embodiment of the present application is shown.

[0036] Figure 2 A second schematic diagram of a diffractive neural network in an embodiment of the present application is shown.

[0037] Figure 3 A schematic structural diagram of a reconfigurable micro-nano unit in an embodiment of the present application is shown.

[0038] Figure 4 A third schematic diagram of a diffractive neural network in an embodiment of the present application is shown.

[0039] Figure 5 A fourth schematic diagram of a diffractive neural network in an embodiment of the present application is shown.

[0040] Figure 6 A first schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0041] Figure 7 A second schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0042] Figure 8 A third schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0043] Fig. 9 A fourth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0044] Fig.10 A fifth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0045] Fig.11 A sixth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown.

[0046] Reference numerals:

[0047] 1- Focusing light control system; 2- Metasurface; 3- Control laser; 4- Reconfigurable micro-nano unit;

[0048] 41-micro-nano structure; 42-spacer structure; 43-substrate; 5-substrate layer;

[0049] 6 - first diffraction layer; 7 - second diffraction layer; 8 - third diffraction layer. DETAILED DESCRIPTION

[0050] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present application will be more comprehensive and complete and the concepts of the example embodiments will be fully conveyed to those skilled in the art. The accompanying drawings are only schematic illustrations of the present application and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted.

[0051] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present application. However, those skilled in the art will appreciate that the technical solution of the present application may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring the present application and making the various aspects of the present application obscure.

[0052] Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processing unit devices and / or microcontroller devices.

[0053] In order to make the diffraction neural network reconfigurable and thus have the ability of dynamic training, the diffraction layer used to provide the phase of the input light in the diffraction neural network needs to have the ability of reconfiguration. In the related art, it is proposed to use phase change materials to form the diffraction layer of the reconfigurable diffraction neural network, so as to achieve the state of the phase change material through external stimulation to achieve the training of the reconfigurable diffraction neural network, or to achieve the switching of the types of tasks that can be processed by the reconfigurable diffraction neural network. In the related art, the regulation of the phase change material is usually to use a high-power microscope objective to focus the control laser on the phase change material of the diffraction layer of the reconfigurable neural network to make the phase change of the phase change material. However, as the optical computing tasks that need to be processed become more and more complex, the number of diffraction layers based on phase change materials in the reconfigurable diffraction neural network is increasing, that is, the number of phase change materials that need to be phase-changed in the reconfigurable diffraction neural network is increasing. The method of using a high-power microscope objective to regulate each phase change material in turn takes a long time, resulting in the slow speed and low efficiency of the reconfigurable diffraction neural network in processing optical computing tasks.

[0054] In order to overcome the above-mentioned defects of the related technology, the present application provides a diffraction neural network that can achieve efficient parallel control of the diffraction layer composed of phase change materials, thereby greatly improving the processing speed of the diffraction neural network for optical computing tasks.

[0055] The diffraction neural network provided in this application includes: an input port, a diffraction system and an output port.

[0056] Specifically, the input port is used to receive input light and guide the input light to the diffraction system. The working band of the input light can be one or more of the visible light band, the near infrared band and the terahertz band, or can be an optical signal of other working bands, which is not limited in this application.

[0057] The diffraction system includes a diffraction layer and a focusing light control system. The diffraction layer is used to perform corresponding optical calculations using input light. The diffraction layer includes a metasurface, which is used to modulate the received input light. That is, in the embodiment of the present application, the metasurface is a physical carrier of neurons in the diffraction neural network. Specifically, the metasurface includes at least two reconfigurable micro-nano units based on phase change materials. The reconfigurable micro-nano units serve as functional units of the diffraction layer, further constituting reconfigurable neurons.

[0058] The focusing light control system is used to control the phase change of the phase change material in the diffraction layer by controlling the laser; specifically, the focusing light control system actually controls the conversion of the phase change material in the reconfigurable micro-nano unit between the crystalline state and the amorphous state to achieve phase change control of the reconfigurable micro-nano unit based on the phase change material. The equivalent refractive index of the reconfigurable micro-nano unit will change with the phase conversion of the phase change material, that is, the phase provided by the reconfigurable micro-nano unit will also change accordingly. Correspondingly, the phase provided by the metasurface composed of the reconfigurable micro-nano units will also change, and the modulation effect of the metasurface on the input light will also change, so that the diffraction layer of the entire diffraction neural network can be reconstructed according to actual task requirements to achieve the task requirements.

[0059] In detail, the focusing light control system can focus high-energy-density incident light (such as laser) onto a light focus that is smaller than or equal to the size of a reconfigurable micro-nano unit in the metasurface, so as to heat the phase change material at this light focus and cause it to undergo a phase change, thereby achieving phase change regulation of the phase change material of the reconfigurable micro-nano unit and thus changing the equivalent refractive index of the micro-nano structure.

[0060] The output port is used to receive the output light from the diffraction layer controlled by the phase change of the focusing light control system, and send the output light to a pre-set detector array, and finally analyze the content detected by the detector array to obtain the final light calculation result.

[0061] In one embodiment, the diffraction layer includes at least two layers of metasurfaces, and the metasurfaces are arranged in sequence along the propagation direction of the optical path. Moreover, all the metasurfaces are arranged on the same side of the base layer of the diffraction neural network. In fact, in the diffraction neural network, the diffraction layer uses the input light to simulate the forward propagation process in the fully connected neural network; according to the Fresnel diffraction theorem, when the input light passes through the diffraction layer, it can be propagated again to the next diffraction layer as a secondary radiation source, that is, the input light completes all network matrix operations during the diffraction process and is modulated into the corresponding output light. It should be noted that the arrangement intervals of the metasurfaces in the diffraction layer can be equal or unequal, and are specifically set according to actual needs and are not limited here.

[0062] As an example, Figure 1 FIG. 1 shows a first schematic diagram of a diffractive neural network in an embodiment of the present application. Figure 1 As shown, in the diffractive neural network provided in one embodiment of the present application, the diffractive layer in the diffractive neural network includes five layers of metasurfaces 2, the direction of the arrow is the propagation direction of the optical path, the five layers of metasurfaces 2 are arranged in sequence along the propagation direction of the optical path, and the five layers of metasurfaces 2 are all arranged on the base layer 5. In the process of optical calculation, the input light is first guided by the input port to the first layer of metasurface along the optical path, and the first layer of metasurface modulates the input light; the input light modulated by the first layer of metasurface continues to propagate along the optical path propagation direction to the second layer of metasurface, and the second layer of metasurface modulates the input light again; the input light modulated by the second layer of metasurface continues to propagate along the optical path propagation direction to the third layer of metasurface, ..., the fifth layer of metasurface modulates the input light modulated by the fourth layer of metasurface to obtain output light; the output light propagates along the optical path propagation direction to the output port, and the output port then emits the output light to the detector array for receiving the output light. Finally, the content detected by the detector array is analyzed to obtain the final optical calculation result.

[0063] Figure 2 A second schematic diagram of a diffractive neural network in an embodiment of the present application is shown. Figure 1 and Figure 2 As shown, in the diffractive neural network, the diffractive layer includes at least two layers of metasurfaces 2, and the two layers of metasurfaces 2 are arranged on the same side of the base layer 5 of the diffractive neural network; each layer of metasurface 2 includes at least two reconfigurable micro-nano units based on phase change materials, therefore, the focusing light control system 1 will control the phase change of each reconfigurable micro-nano unit 4 by controlling the laser 3.

[0064] In one embodiment, the diffraction layer may also include only a single-layer diffraction layer, that is, the diffraction neural network provided by this embodiment is a single-layer neural network.

[0065] In one embodiment, the reconfigurable micro-nano unit 4 includes a substrate 43 , a micro-nano structure 41 and a spacing structure 42 . Figure 3 FIG. 4 shows a schematic diagram of the structure of a reconfigurable micro-nano unit in an embodiment of the present application. Figure 3 As shown, in the reconfigurable micro-nano unit, the micro-nano structure 41 is arranged on one side of the substrate 43 , and the spacing structure 42 is arranged around the micro-nano structure 41 .

[0066] In one embodiment, the material of the micro-nano structure of the reconfigurable micro-nano unit is a phase change material, and accordingly, the spacing structure of the reconfigurable micro-nano unit can be air or other materials that are transparent in the wavelength band of the input light.

[0067] In one embodiment, the material of the spacing structure in the reconfigurable micro-nano unit is a phase change material. Accordingly, the micro-nano structure of the reconfigurable micro-nano unit can be air or other materials that are transparent in the wavelength band of the input light.

[0068] When the focusing light control system performs phase change control on the reconfigurable micro-nano unit, the focusing light control system will focus the control laser on the reconfigurable micro-nano unit. That is, the focusing light control system generates a light focus that matches each reconfigurable micro-nano unit in the diffraction layer one by one, ensuring that the focusing light control system can focus the control laser on the corresponding reconfigurable micro-nano unit, thereby realizing phase change control of all reconfigurable micro-nano units in the diffraction layer.

[0069] In one embodiment, since the diffractive neural network in the present application realizes the regulation of the reconfigurable micro-nano units based on phase change materials in the diffractive layer through the focusing light control system, the diffractive layer is set at one focal length of the focusing light control system. Figure 1 As shown, the focal length of the focusing light control system is f, and the diffraction layer is arranged at one times the focal length of the focusing light control system, that is, each metasurface in the diffraction layer is arranged at one times the focal length of the focusing light control system.

[0070] It should be noted that the one-focal-length position in the embodiment of the present application does not refer to the one-focal-length position in a strict sense, but is within a reasonable tolerance range near the one-focal-length. It can be understood that if the focal length of the focusing light control system is f, the one-focal-length position can be expressed as: f±Δx, where Δx represents a reasonable tolerance. That is to say, in the present application, the one-focal-length position is equivalent to the one-focal-length plus / minus a reasonable tolerance.

[0071] Continuation of the above Figure 1 For example, after obtaining the final optical calculation result, the diffractive neural network can be further trained based on the final optical calculation result, or different optical calculation tasks can be performed based on the diffractive neural network with reconfigurable capabilities. Therefore, by performing phase change control on the diffractive layer in the diffractive neural network according to the focusing optical control system, the following can be achieved, including but not limited to:

[0072] (1) Efficient training of reconfigurable diffractive neural networks;

[0073] (2) Efficient switching of reconfigurable diffractive neural networks for different subtasks in large-scale optical computing tasks.

[0074] During the training process of the reconfigurable diffraction neural network, the input light is irradiated onto the input port and transmitted to the diffraction system through the input port. After the input light is modulated by the diffraction layer in the diffraction system, the output light is obtained. The output light is emitted through the output port to the detector array with the preset settings completed. The content detected by the detector array is analyzed to obtain the result of this optical calculation. Subsequently, according to the result of this optical calculation, a control strategy for the diffraction layer is formulated, and the focusing light control system is controlled according to the control strategy to control each super surface in the diffraction neural network respectively, so that the phase of each super surface in the diffraction layer changes accordingly. After the control is completed, the input light is emitted to the input port again, the result of this optical calculation is recorded again, and the focusing light control system is controlled according to the result of this optical calculation to control each super surface in the diffraction neural network respectively. Repeat the above process until the computing power of the diffraction neural network meets the preset conditions, thereby completing the training of the diffraction neural network. It should be noted that when the computing power of the diffraction neural network reaches the preset conditions, it is not necessary to stop the training of the diffraction neural network, and the diffraction neural network can still be optimized. During this training process, the present application can quickly regulate each metasurface in the diffraction layer through the focusing light control system, which actually greatly shortens the training time of the diffraction neural network, improves the efficiency of the diffraction neural network training, and also makes the subsequent adjustment and optimization of the diffraction neural network more efficient.

[0075] In large-scale optical computing tasks, there are often many different types of optical computing subtasks. The reconfigurable diffraction neural network has the ability to be reconfigured and can achieve changes in the phase of the diffraction layer to adapt to different optical computing subtasks. After the reconfigurable diffraction neural network has completed training for multiple different optical computing subtasks, the control state of each metasurface in the diffraction layer can be recorded when the diffraction neural network is applicable to different optical computing subtasks. For example, if the large-scale optical computing task includes the first subtask, the second subtask, and the third subtask, the three subtasks are different optical computing tasks, and the initial state of the diffraction neural network has been set to a state that can complete the first subtask, then after the diffraction neural network completes the first subtask, it is necessary to control the diffraction layer through the focusing light control system to change the phase of the diffraction layer so that the diffraction neural network can complete the second subtask, and the same is true for the subsequent completion of the third subtask. In the application, the focusing light control system can quickly control all the metasurfaces in the diffraction layer according to the recorded metasurface control states for different optical computing subtasks, so that the diffraction neural network can quickly change the phase of the diffraction layer to adapt to different optical computing tasks, thereby greatly shortening the time to complete large-scale optical computing tasks, improving the computing efficiency of the diffraction neural network, and also increasing the flexibility of the diffraction neural network.

[0076] In one embodiment, the focusing light control system is a multi-focus focusing light control system. In the process of realizing phase change control, the multi-focus focusing light control system uses a multi-focus parallel control method; that is, phase change control is performed on multiple reconfigurable micro-nano units at the same time. Specifically, Figure 1 and Figure 2 The focusing light control system shown is a multi-focus focusing light control system. The multi-focus focusing light control system will focus the control laser on the corresponding reconfigurable micro-nano unit based on phase change materials on the metasurface according to the light focus corresponding to each reconfigurable micro-nano unit, thereby realizing rapid phase change control of the entire metasurface.

[0077] In one embodiment, the multi-focal focusing light control system includes a phase modulation module and a lens module; wherein the phase modulation module is used to modulate the wavefront of the control laser and send the control laser after wavefront modulation to the lens module; the lens module is used to focus the control laser after wavefront modulation to the phase change material of the diffraction layer, that is, to focus the control laser input into the multi-focal focusing light control system so that the control laser can be focused on each reconfigurable micro-nano unit respectively.

[0078] In one embodiment, the phase modulation module is disposed at an entrance pupil of the lens module.

[0079] That is to say, the control laser first passes through the phase modulation module, the phase modulation module modulates the wavefront of the control laser, and transmits the control laser after wavefront modulation to the lens module, the lens module focuses the control laser after wavefront modulation, thereby forming multiple light focal points, all light focal points are in the same plane, and each light focal point corresponds to each reconfigurable micro-nano unit in the diffraction layer, thereby realizing independent regulation of each reconfigurable micro-nano unit, and being able to quickly perform phase change regulation on each reconfigurable micro-nano unit in the diffraction layer, thereby improving the efficiency of diffraction neural network reconstruction.

[0080] In one embodiment, the control laser may be parallel light, and the phase modulation module may modulate the wavefront of the parallel light into a convergent wavefront, and transmit the wavefront-modulated control laser to the lens module.

[0081] In one embodiment, the phase modulation module may be a spatial light modulator (Spatial Light Modulator, SLM), and the spatial light modulator includes a liquid crystal spatial light modulator and also includes a spatial light modulator composed of an adjustable metasurface.

[0082] In one embodiment, the phase modulation module may also be a digital micromirror device (DMD). DMD is a device based on semiconductor manufacturing technology and composed of a high-speed digital optical reflective switch array.

[0083] In one embodiment, the lens module includes at least one superlens or at least one microlens. That is, in the embodiment of the present application, the lens module can be composed of a superlens, a superlens group, a superlens group array, or a microlens, a microlens group, or a microlens array.

[0084] When the lens module is a superlens array, there is a one-to-one correspondence between each superlens in the superlens array and each reconfigurable micro-nano unit in the diffraction layer, that is, each superlens in the superlens array is used to perform phase change regulation for the corresponding reconfigurable micro-nano unit, and each superlens controls the laser to focus on the corresponding reconfigurable micro-nano unit.

[0085] When the lens module is a microlens array, similar to the superlens array, each microlens in the microlens array has a one-to-one correspondence with each reconfigurable micro-nano unit in the diffraction layer, that is, each microlens in the microlens array is used to perform phase change regulation for the corresponding reconfigurable micro-nano unit, and each microlens controls the laser to focus on the corresponding reconfigurable unit.

[0086] In one embodiment, the lens module includes at least one super lens and at least one micro lens. That is, in another embodiment of the present application, the lens module can also be an optical system composed of a micro lens and a super lens.

[0087] It should be noted that the lens module is an optical system with a large numerical aperture, which can produce light focal points with a spacing of hundreds of nanometers. The large numerical aperture ensures that the light focal points are small and the energy is concentrated, which is more conducive to precise pixel-level control.

[0088] It should be noted that if the lens module includes a superlens, the focusing phase of the superlens must satisfy:

[0089]

[0090] Wherein, φ(x, y) is the focusing phase of the metalens, (x, y) is the position on the metalens, λ represents the operating band of the controlled laser, and f is the focal length of the metalens. In this application, the focusing phase of all metalenses satisfies this formula.

[0091] In one embodiment of the present application, the light source used to emit control laser to the multi-focal focusing light control system can be a single light source; that is, the corresponding phase change regulation can be achieved by switching the light source. When the light source is turned on, the multi-focal focusing light control system performs phase change regulation on the reconfigurable micro-nano unit in the diffraction layer of the diffraction neural network. In addition, the turn-on time of the light source can be controlled to control the regulation time of the reconfigurable micro-nano unit, so as to achieve the actual required crystallization ratio of the phase change material in the reconfigurable micro-nano unit, and complete the precise regulation of the equivalent refractive index of the reconfigurable micro-nano unit.

[0092] In another embodiment of the present application, when a superlens array or a microlens array is used as a lens module, the light source for emitting the control laser can be a light source array composed of multiple sub-light sources; it should be noted that each sub-light source in the light source array has a one-to-one correspondence with each superlens in the superlens array, or each sub-light source in the light source array has a one-to-one correspondence with each microlens in the microlens array. In this embodiment, the switch of each sub-light source can be controlled separately, so as to realize the independent phase change regulation of a reconfigurable micro-nano unit in the diffraction layer, that is, by controlling whether each sub-light source is turned on or off, the control of the corresponding reconfigurable micro-nano unit can be realized, so as to realize the independent regulation of each reconfigurable micro-nano unit in the diffraction layer of the diffraction neural network. Moreover, by setting a different start-up time for each sub-light source, the regulation time of each reconfigurable micro-nano unit can be realized, so that the phase change material in each reconfigurable micro-nano unit can reach different crystallization ratios, that is, each reconfigurable micro-nano unit after regulation can have a different equivalent refractive index, which can better meet the actual reconfigurable needs, thereby realizing high-precision reconstruction of the diffraction neural network and greatly enhancing the flexibility of the diffraction neural network reconstruction.

[0093] In one embodiment, the focusing light control system can also be a single-focus focusing light control system. As an example, the single-focus focusing light control system includes a laser scanning module and a lens module; the laser scanning module is used to deflect the control laser; the lens module is used to focus the control laser. That is to say, along the optical path direction of the control laser, the arrangement order of the laser scanning module and the lens module is not unique, and the laser scanning module can be arranged upstream of the lens module, or the lens module can be arranged upstream of the laser module.

[0094] Along the optical path direction of the control laser, when the laser scanning module is arranged upstream of the lens module, the laser scanning module is used to deflect the control laser and transmit the deflected control laser to the lens module, and the lens module is used to focus the deflected control laser onto the phase change material of the diffraction layer; when the lens module is arranged upstream of the laser scanning module, the lens module first converges the control laser, and then the laser scanning module deflects the control laser converged by the lens module.

[0095] Specifically, Figure 4 A third schematic diagram of a diffractive neural network in an embodiment of the present application is shown, Figure 5 A fourth schematic diagram of a diffractive neural network in an embodiment of the present application is shown. Figure 4 and Figure 5 As shown, Figure 4 and Figure 5 The focusing light control system 1 in the embodiment is a single-focus focusing light control system.

[0096] Unlike the multi-focus optical control system, the single-focus optical control system uses a single-focus fast scanning method to achieve phase change control; that is, each reconfigurable micro-nano unit is phase-controlled in turn. Specifically, Figure 4 The multiple control laser beams in the system are emitted in a certain order by the single-focus focusing optical control system. The single-focus focusing optical control system will deflect the control laser in a certain order through the laser scanning module according to the optical focus corresponding to each reconfigurable micro-nano unit, and the lens module will focus the deflected control laser to the corresponding reconfigurable unit, and perform phase change control on each reconfigurable micro-nano unit in turn, thereby realizing rapid control of the entire metasurface. It should be noted that Figure 4 The dotted part in the figure represents the control laser emitted by the single-focus focusing light control system for the reconfigurable micro-nano units of different metasurfaces.

[0097] Figure 5 Only the diffraction layer composed of a single-layer metasurface is shown, and the diffraction layer is arranged on the base layer of the diffraction neural network. Figure 5 The input port and the output port provided on the base layer at the same time are not shown. Figure 5 In the single-focus focusing light control system, the laser is controlled to be focused on the reconfigurable micro-nano unit through the single-focus focusing light control system, thereby realizing the phase change regulation of the reconfigurable micro-nano unit. After the phase change regulation of this reconfigurable micro-nano unit is completed, the single-focus focusing light control system will deflect the control laser through the laser scanning module, and through the cooperation with the lens module, the control laser is focused on another reconfigurable micro-nano unit that needs phase change regulation. Repeat the above process until the phase change regulation of all reconfigurable micro-nano units is completed.

[0098] It should be noted that in the present application, the focusing light control system can achieve phase change regulation by controlling the laser to perform photothermal effect, or by controlling the laser to achieve phase change regulation by laser pulse irradiation, which is not limited in the present application.

[0099] It should be noted that in the control process of the above-mentioned single-focus focusing light control system, the phase change control of each reconfigurable micro-nano unit is quickly performed by means of rapid scanning of the single focus of light, which means that the actual optical computing task requirements can be met and the phase change control that meets the requirements can be completed within the task limit time. In other words, the control speed of the single-focus focusing light control system in this application can be dynamically adjusted according to the time required to complete the phase change control and the number of reconfigurable micro-nano units that need to be controlled in the optical computing task, so as to meet the target requirements of the optical computing task. As an example, if in the training process of a certain diffraction neural network, the total number of reconfigurable micro-nano units in the diffraction layer is 500, and the total time of the given training process is 1 second, then the single-focus focusing light control system needs to complete the phase change control of 500 reconfigurable micro-nano units within 1 second, that is, for each reconfigurable micro-nano unit, the single-focus focusing light control system needs to complete the deflection of the control laser within 1 / 500 second.

[0100] The diffraction neural network in the present application can be regarded as a two-dimensional diffraction neural network, and multiple diffraction neural networks can constitute a three-dimensional diffraction neural network; the three-dimensional diffraction neural network can perform multiple tasks in parallel. For the reconfigurable three-dimensional diffraction neural network, the problem caused by executing multiple tasks in parallel is that when the types of tasks executed in parallel by the three-dimensional diffraction neural network change, all diffraction layers of the three-dimensional diffraction neural network need to be reconstructed. Since there are multiple diffraction layers in the three-dimensional diffraction neural network, and each diffraction layer has multiple layers of metasurfaces, the existing technology makes it difficult to independently, flexibly and efficiently phase-change control the multiple diffraction layers in the three-dimensional diffraction neural network.

[0101] To this end, the present application proposes a three-dimensional diffraction neural network, which includes a three-dimensional input port, a three-dimensional diffraction system and a three-dimensional output port.

[0102] Specifically, the three-dimensional diffraction neural network can be formed by vertically stacking and packaging multiple diffraction neural networks. Therefore, the three-dimensional input port includes at least two of the above-mentioned input ports. The three-dimensional diffraction system includes at least two of the above-mentioned diffraction layers; the three-dimensional diffraction system also includes at least two of the above-mentioned focusing light control systems; and the three-dimensional output port includes at least two of the above-mentioned output ports.

[0103] Since the three-dimensional diffraction neural network is formed by vertically stacking and packaging multiple diffraction neural networks, the diffraction layer part of the three-dimensional diffraction neural network is formed by vertically stacking and packaging at least two of the above-mentioned diffraction layers.

[0104] In one embodiment, the focusing light control system is matched one-to-one with the diffraction layer; that is, in the three-dimensional diffraction neural network, each diffraction layer has a corresponding focusing light control system, and each focusing light control system can perform phase change control on the corresponding diffraction layer.

[0105] Moreover, each diffraction layer is arranged at one focal length of the corresponding focusing light control system. Similar to the two-dimensional diffraction neural network, one focal length here is not strictly equal to one focal length, but equal to one focal length plus / minus a reasonable tolerance.

[0106] It should be noted that in the three-dimensional diffraction neural network, the input ports in the three-dimensional input port, the diffraction layers, and the output ports in the three-dimensional output port are matched one by one. That is to say, in the process of the three-dimensional diffraction neural network executing multiple tasks in parallel, each input port in the three-dimensional input port guides the input light to the corresponding diffraction layer, and the corresponding output port then guides the output light modulated by the corresponding diffraction layer to the detector array, and analyzes the content detected by the detector, thereby obtaining the calculation results of multiple tasks.

[0107] In the three-dimensional diffractive neural network provided in the present application, similar to the two-dimensional diffractive neural network, the focusing light control system can be a multi-focus focusing light control system or a single-focus focusing light control system.

[0108] In one embodiment, when the focal lengths of all the focusing light control systems are the same, all the focusing light control systems are not located in the same plane.

[0109] Specifically, Figure 6 A first schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown. Figure 6 As shown, Figure 6 Only the three-dimensional diffraction system part of the three-dimensional diffraction neural network in this embodiment is shown. In this embodiment, the three-dimensional diffraction system includes a first diffraction layer 6, a second diffraction layer 7 and a third diffraction layer 8, and the focusing light control system 1 is a multi-focus focusing light control system. The first diffraction layer, the second diffraction layer and the third diffraction layer respectively have corresponding multi-focus focusing light control systems, and the first diffraction layer, the second diffraction layer and the third diffraction layer are respectively arranged at one focal length of the corresponding multi-focus focusing light control system, f 1 、f 2 and f 3 are the focal lengths of the multi-focal focusing light control systems corresponding to the first diffraction layer, the second diffraction layer, and the third diffraction layer, respectively. The first diffraction layer, the second diffraction layer, and the third diffraction layer are phase-changed and regulated by the corresponding multi-focal focusing light control systems according to actual needs, thereby realizing independent and efficient phase-change regulation of multiple diffraction layers in the three-dimensional diffraction neural network. In this embodiment, the focal lengths of the multi-focal focusing light control systems corresponding to each diffraction layer are equal, that is, f 1 =f 2 =f 3 Since the first diffraction layer, the second diffraction layer and the third diffraction layer are located in different planes, the multi-focus focusing light control system corresponding to each diffraction layer is also located in a different plane. Figure 6In the figure, the dotted part in the focusing light control system 1 is used to indicate that the focusing light control system corresponding to the diffraction layer is a whole, the focusing light control systems at both ends of the dotted part are the same system, and the dotted part can be air or a transparent material in the wavelength band of the controlled laser, as long as the multi-focus focusing light control systems corresponding to the diffraction layer are located in the same plane.

[0110] Figure 7 A second schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown. Figure 7 As shown, Figure 7 The focusing light control system 1 in the embodiment is a single-focus focusing light control system. 1 、f 2 and f 3 are the focal lengths of the single-focus focusing light control systems corresponding to the first diffraction layer 6, the second diffraction layer 7, and the third diffraction layer 8, respectively, and the focal lengths of the single-focus focusing light control systems corresponding to each diffraction layer are equal, that is, f 1 =f 2 =f 3 Since the first diffraction layer, the second diffraction layer and the third diffraction layer are located in different planes, the single-focus focusing light control system corresponding to each diffraction layer is also located in a different plane. It should be noted that Figure 7 The dotted part in the figure represents the control laser emitted by the single-focus focusing light control system for the reconfigurable micro-nano units of different metasurfaces.

[0111] In one embodiment, when all the focusing light control systems are located in the same plane, the focal lengths of all the focusing light control systems are different.

[0112] Specifically, Figure 8 A third schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown. Fig. 9 FIG. 4 shows a fourth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application. Figure 8 and Fig. 9 As shown, in this embodiment, the three-dimensional diffraction system includes a first diffraction layer 6, a second diffraction layer 7 and a third diffraction layer 8, all focusing light control systems are located in the same plane, and the focusing light control system is a multi-focal focusing light control system. A is the multi-focal focusing light control system corresponding to the first diffraction layer, B is the multi-focal focusing light control system corresponding to the second diffraction layer, and C is the multi-focal focusing light control system corresponding to the third diffraction layer. A, B and C are all in the same plane. It should be understood that since the distances from the plane where the first diffraction layer, the second diffraction layer and the third diffraction layer are located to the plane where the multi-focal focusing light control system is located are different, and each diffraction layer is set at one times the focal length of the corresponding multi-focal focusing light control system, the focal lengths of the three multi-focal focusing light controls A, B and C are different, that is, f 1 ≠f 2 ≠f 3, so that A, B and C can respectively produce light foci in different planes, realizing the phase change regulation of all reconfigurable micro-nano units in the three-dimensional diffractive neural network.

[0113] Fig.10 A fifth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown. Fig.10 As shown, in this embodiment, the three-dimensional diffraction system includes a first diffraction layer 6, a second diffraction layer 7 and a third diffraction layer 8, all focusing light control systems 1 are located in the same plane, and the focusing light control system is a single-focus focusing light control system. Figure 8 and Fig. 9 The embodiment shown is similar, in Fig.10 In the corresponding embodiment, f 1 ≠f 2 ≠f 3 , so that each single-focus focusing light control system can perform phase change control on the corresponding diffraction layer. It should be noted that Fig.10 The dotted part in the figure represents the control laser emitted by the single-focus focusing light control system for the reconfigurable micro-nano units of different metasurfaces.

[0114] Furthermore, when each focusing light control system is located in a different plane, the focal length of each focusing light control system may also be different.

[0115] It should be emphasized that since the diffraction layer part in the three-dimensional diffraction system is composed of multiple diffraction layers stacked vertically, and the above Figures 6 to 10 In the embodiment shown, each diffraction layer needs to be phase-changed. Therefore, if the diffraction layer is to be phase-changed through the focusing light control system corresponding to each diffraction layer, it is necessary to ensure that the control laser output by each focusing light control system will not be blocked, so that it can be effectively focused on the target reconfigurable micro-nano unit that needs phase-change control.

[0116] In one embodiment, the metasurface to be regulated can be staggered in the direction of vertical stacking of the diffraction layers, and the base layer of the diffraction neural network should be a transparent material in the wavelength band of the control laser, so that the base layer can at least transmit the control light, so that each focusing light control system can focus the control laser on the reconfigurable micro-nano unit of the corresponding diffraction layer, thereby realizing the phase change regulation of the entire three-dimensional diffraction neural network. Specifically, when the focusing light control system is a multi-focus focusing light control system, such as Figure 6 and Fig. 9 As shown, Figure 6 and Fig. 9 All metasurfaces in the paper are in a staggered configuration.

[0117] In addition, not all metasurfaces need to be staggered. Figure 7 and Fig.10 As shown, due to Figure 7 and Fig.10 The positional relationship of the single-focus focusing light control system in Figure 7 and Fig.10 Only the super surface on the left may be blocked, so it needs to be set up in a staggered manner; Figure 7 and Fig.10 There is no occlusion on the hypersurface on the middle right, so no misalignment setting is required.

[0118] Similarly, in a three-dimensional diffraction neural network, there is a situation where not all diffraction layers need to be phase-changed and regulated. In this case, the metasurface of the diffraction layer that does not need to be regulated may not be staggered. Fig.11 A sixth schematic diagram of a three-dimensional diffractive neural network according to an embodiment of the present application is shown. Fig.11 As shown, in this embodiment, the first diffraction layer 6 and the second diffraction layer 7 need to be phase-changed, and the third diffraction layer 8 does not need to be phase-changed, and the focusing light control system in this embodiment is a multi-focus focusing light control system. Therefore, only the first diffraction layer and the second diffraction layer have corresponding multi-focus focusing light control systems A and B, and the metasurfaces in the first diffraction layer and the second diffraction layer are both staggered. The metasurface in the third diffraction layer does not need to be staggered with the metasurfaces in the first diffraction layer and the second diffraction layer.

[0119] That is to say, in the three-dimensional diffraction neural network of the present application, no matter whether the focusing light control system is a multi-focus focusing light control system or a single-focus focusing light control system, no matter how many diffraction layers need to be phase-changed, the positional relationship between the metasurfaces only needs to satisfy the following: each metasurface that needs to be phase-changed can be phase-changed by the corresponding focusing light control system.

[0120] In another embodiment, a control laser of a different wavelength can be set for the focusing light control system corresponding to each diffraction layer to avoid the problem that the control laser focused on the target reconfigurable micro-nano unit is blocked during the phase change control process, resulting in the target reconfigurable micro-nano unit being unable to perform phase change control.

[0121] Figures 6 to 11 The three-dimensional diffraction neural network shown is only an example and does not represent the actual setting method of the diffraction layer part and the focusing light control system in the three-dimensional diffraction neural network of the present application. That is, the number of diffraction layers and the number of metasurfaces in the three-dimensional diffraction neural network in the present application are not limited.

[0122] As an example, Fig. 9In the three-dimensional diffraction neural network shown, the diffraction layer part includes the first diffraction layer, the second diffraction layer and the third diffraction layer. Assume that the three diffraction layers perform different tasks, the first diffraction layer needs to perform the first subtask, the second diffraction layer needs to perform the second subtask, and the third diffraction layer needs to perform the third subtask. For example:

[0123] (1) Before the three diffraction layers perform the three corresponding tasks respectively, the three diffraction layers need to be trained respectively. Since each diffraction layer has a corresponding multi-focus focusing light control system, each diffraction layer can be efficiently trained at the same time, which greatly improves the training efficiency of the three-dimensional diffraction neural network and shortens the overall time for the three-dimensional diffraction neural network to complete the optical computing task.

[0124] (2) After the training of the three diffraction layers is completed, if there is a diffraction layer that still needs to be adjusted and optimized, the diffraction layer can be adjusted and optimized again through the corresponding multi-focus focusing light control system, so that the regulation of the diffraction layer part of the three-dimensional diffraction neural network is more targeted, thereby improving the flexibility of the reconstruction of the three-dimensional diffraction neural network.

[0125] (3) If the first diffraction layer needs to change the optical computing task to be executed from the original first subtask to the second subtask after executing the first subtask, the first diffraction layer can be phase-changed by the multi-focus focusing optical control system corresponding to the first diffraction layer, so that the phase of the first diffraction layer can be quickly changed, thereby processing the second subtask in time and improving the overall efficiency of the optical computing task.

[0126] In summary, the three-dimensional diffraction neural network provided in the present application performs phase change regulation on different diffraction layers through multiple focusing light control systems, thereby improving the training efficiency of the three-dimensional diffraction neural network, increasing the flexibility of the reconstruction of the three-dimensional diffraction neural network, and improving the overall efficiency of optical computing tasks.

[0127] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the appended claims.

Claims

1. A diffractive neural network, characterized in that: The diffractive neural network comprises: an input port, a diffractive system and an output port; Wherein, the input port is used to receive input light and guide the input light to the diffraction system; The diffraction system includes a diffraction layer and a focusing light control system; the diffraction layer includes a metasurface, and the metasurface is used to modulate the input light; the metasurface includes at least two reconfigurable micro-nano units based on phase change materials; The focusing light control system is used to control the phase change of the phase change material in the diffraction layer by controlling the laser; The output port is used to receive the output light from the diffraction layer that is phase-changed and regulated by the focusing light control system, and emit the output light to a detector array for receiving the output light.

2. The diffractive neural network according to claim 1, characterized in that The diffraction layer includes at least two layers of metasurfaces, and the metasurfaces are sequentially arranged at intervals along the propagation direction of the light path.

3. The diffractive neural network according to claim 1, characterized in that The reconfigurable micro-nano unit comprises a substrate, a micro-nano structure and a spacer structure; Wherein, the material of the micro-nano structure is a phase change material.

4. The diffractive neural network according to claim 1, characterized in that The reconfigurable micro-nano unit comprises a substrate, a micro-nano structure and a spacer structure; Wherein, the material of the spacing structure is phase change material.

5. The diffractive neural network according to claim 1, characterized in that: The diffraction layer is arranged at one focal length of the focusing light control system.

6. The diffractive neural network according to claim 1, characterized in that: The focusing light control system is a multi-focus focusing light control system or a single-focus focusing light control system.

7. The diffractive neural network according to claim 6, characterized in that: The multi-focal point focusing light control system comprises a phase modulation module and a lens module; The phase modulation module is used to modulate the wavefront of the control laser and transmit the control laser after wavefront modulation to the lens module; The lens module is used to focus the control laser after wavefront modulation onto the phase change material of the diffraction layer.

8. The diffractive neural network according to claim 7, characterized in that: The phase modulation module is a spatial light modulator or a digital micromirror device; The phase modulation module is arranged at the entrance pupil of the lens module.

9. A three-dimensional diffraction neural network, characterized in that: The three-dimensional diffraction neural network includes a three-dimensional input port, a three-dimensional diffraction system and a three-dimensional output port; Wherein, the three-dimensional input port comprises at least two input ports according to any one of claims 1 to 8; The three-dimensional diffraction system comprises at least two diffraction layers according to any one of claims 1 to 8; The three-dimensional diffraction system further comprises at least two focusing light control systems as described in any one of claims 1 to 8; The three-dimensional output port comprises at least two output ports as described in any one of claims 1-8.

10. The three-dimensional diffractive neural network according to claim 9, characterized in that: The focusing light control system and the diffraction layer are matched one by one; The diffraction layer is arranged at one focal length of the focusing light control system.

Citation Information

Cited By

  • D2NNs-based terahertz phased array device and implementation method thereof

    CN119200292A

  • A D-based 2 Terahertz phased array device for NNs and implementation method thereof

    CN119200292B