Method, device and program product for adjusting structural parameters of micro-nano structures

Abstract

The invention relates to a method, a device and a computer program product for adjusting structure parameters of a micro-nano structure. The method comprises the following steps: dividing a micro-nano structure into a plurality of layers, wherein each layer in the plurality of layers has a corresponding structure parameter; determining the actual photon response of the micro-nano structure to incident light, wherein the incident light sequentially passes through the plurality of layers along the transmission direction; determining a loss function representing an error of the actual photon response relative to the predetermined photon response; according to the loss function, obtaining a derivative of the loss function relative to the structural parameter of each layer by using a back propagation mode relative to the transmission direction; and for each of the plurality of layers, using the derivative to adjust the structure parameter such that the adjusted micro-nano structure has a predetermined photon response for incident light. According to the invention, the structure parameters of the micro-nano structure can be efficiently and accurately adjusted.

Description

Technical Field

[0001] Embodiments of the present invention generally relate to the field of optics, and more particularly to a method, apparatus, and computer program product for adjusting structural parameters of a micro-nanostructure. Background Art

[0002] With the development of science and technology, micro-nanostructures have been applied to an increasing number of fields. For different application areas, micro-nanostructures with different photon responses may be desired. Therefore, adjusting the structural parameters of micro-nanostructures to achieve the desired photon response is crucial for their application. Typically, during the design of micro-nanostructures, the structural parameters of the micro-nanostructure are adjusted by performing a differential approximation of the actual photon response of the micro-nanostructure relative to its structural parameters to achieve the desired photon response. However, this adjustment method has low accuracy and efficiency. Summary of the Invention

[0003] The present invention provides a method, device and computer program product for adjusting structural parameters of a micro-nano structure, which can efficiently and accurately adjust the structural parameters of the micro-nano structure.

[0004] According to a first aspect of the present invention, a method for adjusting structural parameters of a micro-nanostructure is provided. The method includes: dividing the micro-nanostructure into multiple layers, each layer having corresponding structural parameters; determining an actual photon response of the micro-nanostructure to incident light, the incident light sequentially passing through the multiple layers along a transmission direction; determining a loss function representing the error of the actual photon response relative to a predetermined photon response; obtaining, based on the loss function, a derivative of the loss function with respect to the structural parameters of each layer using a backpropagation method with respect to the transmission direction; and adjusting the structural parameters of each layer using the derivatives so that the adjusted micro-nanostructure has the predetermined photon response to the incident light.

[0005] In some embodiments, determining the actual photon response of the micro-nanostructure to incident light includes: using structural parameters to determine an eigenvector for each layer in a plurality of layers, wherein the elements in the eigenvector represent the eigenmodes of the corresponding layer; for every two adjacent layers in the plurality of layers, determining an interlayer transmission matrix for incident light between every two adjacent layers through electromagnetic field boundary conditions; and determining the actual photon response through all eigenvectors and all interlayer transmission matrices of the plurality of layers.

[0006] In some embodiments, the eigenvector includes a first eigenvector and a second eigenvector, the elements of the first eigenvector represent the eigenmode at the first surface inside the corresponding layer, the elements of the second eigenvector represent the eigenmode at the second surface inside the corresponding layer, and the first surface and the second surface are facing surfaces, and the incident light passes through the first surface and the second surface in sequence along the transmission direction.

[0007] In some embodiments, the first eigenvector has a first correlation with the second eigenvector, the first correlation is represented by the intra-layer transmission matrix of the corresponding layer, the second eigenvector of the first layer in every two adjacent layers of the multiple layers has a second correlation with the first eigenvector of the second layer, the second correlation is represented by the corresponding inter-layer transmission matrix, and the incident light passes through the first layer and the second layer in sequence along the transmission direction.

[0008] In some embodiments, according to the loss function, using the back propagation method relative to the transmission direction, obtaining the derivative of the loss function with respect to the structural parameters of each layer includes: along the direction opposite to the propagation direction, according to the chain rule, through the first association and the second association, successively determining the first derivative of the loss function with respect to the first eigenvector of each layer and the second derivative with respect to the second eigenvector of each layer; and for each layer, using the first derivative and the second derivative, determining the derivative of the loss function with respect to the structural parameters.

[0009] In some embodiments, for each layer in a plurality of layers, using derivatives to adjust structural parameters so that the adjusted micro-nanostructure has a predetermined photon response to incident light includes: for each layer in a plurality of layers, using derivatives to adjust structural parameters to obtain a currently adjusted micro-nanostructure; determining a current actual photon response of the currently adjusted micro-nanostructure to the incident light; determining whether a current error of the current actual photon response relative to the predetermined photon response is within a predetermined error range; and in response to determining that the current error is within the predetermined error range, determining that the currently adjusted micro-nanostructure has the predetermined photon response to the incident light.

[0010] In some embodiments, for each layer in a plurality of layers, using derivatives to adjust structural parameters so that the adjusted micro-nanostructure has a predetermined photon response to incident light also includes: in response to determining that the current error is not within a predetermined error range, determining a current loss function representing the current error; and using the current loss function as a loss function to re-adjust the current structural parameters of each layer in the plurality of layers until the adjusted micro-nanostructure has a predetermined photon response to incident light.

[0011] In some embodiments, the photon response includes at least one of: reflectivity, transmittance, phase shift, and polarization.

[0012] In some embodiments, the multiple layers are virtually divided layers, and the micro-nano structure is an integrated structure.

[0013] In some embodiments, the micro-nanostructure is used for at least one of the following: a grating, an optical film, a metasurface, a reflector, a polarizer, and a sensor.

[0014] In some embodiments, the structural parameter of each layer of the plurality of layers includes at least one of: thickness, vertex coordinates of each vertex, and length of each edge.

[0015] In some embodiments, the loss function represents a mean square error between an actual photon response and a predetermined photon response.

[0016] According to a second aspect of the present invention, a device for adjusting structural parameters of a micro-nanostructure is provided. The device includes: a first unit configured to divide the micro-nanostructure into multiple layers, each layer having corresponding structural parameters; a second unit configured to determine an actual photon response of the micro-nanostructure to incident light, the incident light sequentially passing through the multiple layers along a transmission direction; a third unit configured to determine a loss function representing an error of the actual photon response relative to a predetermined photon response; a fourth unit configured to obtain, based on the loss function, a derivative of the loss function with respect to the structural parameters of each layer using a back-propagation method with respect to the transmission direction; and a fifth unit configured to use the derivative to adjust the structural parameters of each layer in the multiple layers so that the adjusted micro-nanostructure has a predetermined photon response to the incident light.

[0017] According to a third aspect of the present invention, a computer program product is provided, comprising a computer program, which, when executed by a machine, performs the method for adjusting structural parameters of a micro-nanostructure according to the present invention.

[0018] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements.

[0020] Figure 1 A schematic diagram illustrating an example of a micro-nano structure according to an embodiment of the present invention is shown.

[0021] Figure 2 A flow chart of a method for adjusting structural parameters of a micro-nanostructure according to an embodiment of the present invention is illustrated.

[0022] Figure 3 FIG2 shows a schematic diagram of a micro-nano structure divided into multiple layers according to an embodiment of the present invention.

[0023] Figure 4 A flow chart illustrating a process of determining an actual photon response of a micro-nanostructure to incident light according to an embodiment of the present invention is shown.

[0024] Figure 5 is a schematic diagram illustrating an eigenmode of each of a plurality of layers according to an embodiment of the present invention.

[0025] Figure 6 is a schematic diagram illustrating the similarity between the micro-nanostructure and the neural network model according to an embodiment of the present invention.

[0026] Figure 7 is another schematic diagram illustrating the similarity between the micro-nanostructure and the neural network model according to an embodiment of the present invention.

[0027] Figure 8 is a flowchart illustrating a process of obtaining a derivative of a loss function with respect to a structural parameter of each layer according to an embodiment of the present invention.

[0028] Figure 9 is a flow chart illustrating a process of adjusting structural parameters using derivatives according to an embodiment of the present invention.

[0029] Figure 10 A block diagram of an apparatus for adjusting structural parameters of a micro-nanostructure according to an embodiment of the present invention is shown.

[0030] Figure 11 A schematic block diagram of an example device suitable for implementing embodiments of the present disclosure is shown.

[0031] In the various drawings, the same or corresponding reference numerals denote the same or corresponding parts. DETAILED DESCRIPTION

[0032] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to make the present invention more thorough and complete and to fully convey the scope of the present invention to those skilled in the art.

[0033] As used herein, the term "including" and its variations represent open inclusion, i.e., "including but not limited to." Unless otherwise stated, the term "or" means "and / or." The term "based on" means "based at least in part on." The terms "one example embodiment" and "an embodiment" mean "at least one example embodiment." The term "another embodiment" means "at least one additional embodiment." The terms "first," "second," etc. may refer to different or identical objects.

[0034] As described above, micro-nanostructures with different photon responses may be desired for different application areas. Therefore, after a micro-nanostructure is artificially developed or discovered in nature, its structural parameters need to be designed (i.e., adjusted) to achieve the desired photon response. However, the conventional method of adjusting structural parameters through differential approximation has low accuracy and low adjustment efficiency.

[0035] To at least partially address one or more of the aforementioned and other potential issues, exemplary embodiments of the present invention provide a method for adjusting the structural parameters of a micro-nanostructure. According to the method of the disclosed embodiments, the structural parameters of the micro-nanostructure can be adjusted using a loss function through backpropagation, thereby improving the accuracy and efficiency of the adjustment. Thus, a micro-nanostructure with a desired (predetermined) photon response can be obtained.

[0036] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Figure 1 FIG. 1 is a schematic diagram illustrating an example of a micro-nanostructure 100 according to an embodiment of the present invention. Figure 1 As shown, the micro-nanostructure 100 can be a micro-nanostructure artificially developed or found in nature, and it may be an integrated structure with an irregular shape. The micro-nanostructure 100 can be configured to have a desired (predetermined) photon response for use in different fields. In some embodiments, the photon response can include at least one of the following: reflectivity, transmittance, phase shift, and polarization. In some embodiments, the micro-nanostructure with a predetermined photon response can be used for at least one of the following: a grating, an optical film, a metasurface, a reflector, a polarizer, and a sensor.

[0037] The predetermined photon response of the micro-nanostructure 100 can be achieved by adjusting the structural parameters of the micro-nanostructure. Figure 1 The micro-nano structure 100 is taken as an example to illustrate the method for adjusting the structural parameters of the micro-nano structure according to an embodiment of the present invention. Figure 2 A flow chart of a method 200 for adjusting structural parameters of a micro-nanostructure 100 according to an embodiment of the present invention is shown. Figure 2 The method 200 shown may be executed by any controller, processing unit, electronic terminal, or the like.

[0038] like Figure 2 As shown, at block 202, the micro-nano structure 100 is divided into a plurality of layers, each of which has corresponding structural parameters. Figure 1 As shown, the micro-nanostructure 100 has an irregular structural shape. Therefore, when adjusting the structural parameters of the micro-nanostructure 100 to achieve a predetermined photon response, if the micro-nanostructure 100 is adjusted as a whole, the adjustment accuracy will be low. According to an embodiment of the present invention, by layering the micro-nanostructure 100, the micro-nanostructure 100 is regarded as a superposition of multiple layers, and the structural parameters of each layer are adjusted layer by layer, which can greatly improve the adjustment accuracy. Here, the multiple layers are virtual layers, and the micro-nanostructure 100 is actually a one-piece structure.

[0039] Figure 3 FIG2 shows a schematic diagram of a micro-nano structure 100 divided into multiple layers according to an embodiment of the present invention. Figure 3 As shown, the micro-nanostructure 100 can be divided into N layers 300_1 to 300_N, where N is an integer greater than 1. Each of the N layers 300_1 to 300_N can have corresponding structural parameters. In some embodiments, the structural parameters of each of the N layers 300_1 to 300_N can include at least one of the following: thickness, vertex coordinates of each vertex, and length of each edge. Figure 3 The direction indicated by the arrow A represents the transmission direction of the incident light on the micro-nanostructure 100 in the micro-nanostructure.

[0040] Return to reference Figure 2 At block 204, the actual photon response of the micro-nanostructure 100 to the incident light is determined. The incident light is transmitted along a transmission direction (eg, Figure 3 The dielectric constant of the micro-nanostructure 100 may be periodic in at least one direction. In this case, the actual photon response of the micro-nanostructure 100 to incident light can be obtained, for example, using a Rigorous Coupled-Wave Analysis (RCWA) algorithm based on the initial structural parameters of each layer. The RCWA algorithm is an algorithm that calculates the propagation characteristics (e.g., diffraction characteristics) of electromagnetic waves in periodic media based on the Fourier series expansion principle and electrodynamics. Therefore, it is suitable for calculating the actual photon response of a micro-nanostructure to incident light.

[0041] After obtaining the actual photon response of the micro-nanostructure 100 to the incident light, at block 206, a loss function (e.g., a function value of the loss function) is determined that represents the error of the actual photon response relative to a predetermined photon response. The predetermined photon response may be a desired photon response for the micro-nanostructure 100. In some embodiments, the loss function may represent the mean squared error between the actual photon response and the predetermined photon response. For example, the loss function may be similar to the loss function used in the training process of a neural network model.

[0042] At block 208, the derivative of the loss function with respect to the structural parameters of each layer is obtained based on the loss function using a backpropagation method relative to the transmission direction. This derivative can represent the contribution of each layer to the above-mentioned error, thereby accurately adjusting the structural parameters of each layer. Conventionally, when obtaining the actual photon response of the micro-nanostructure to the incident light using a rigorous coupled wave algorithm, it is usually necessary to obtain the above-mentioned contribution by using a differential approximation solution for each structural parameter of each layer. However, the differential approximation solution requires separate calculations for each structural parameter, resulting in a very large amount of calculations and low computational efficiency. In addition, due to limited computing resources, this large amount of calculations may make it difficult to divide the micro-nanostructure 100 into a large number of layers, which further limits the accuracy of adjusting the structural parameters of the micro-nanostructure (and thus setting its photon response).

[0043] In an embodiment of the present invention, the above problem is at least solved by using a back propagation method to obtain the derivative representing the contribution degree. For example, the back propagation method can be used along the same Figure 3 Starting from obtaining the derivative of the loss function with respect to the last layer 300_N, the derivative is propagated forward layer by layer according to the chain rule, thereby accurately and efficiently obtaining the derivative of the loss function with respect to each layer 300_N to 300_1 in sequence (as further described in the example below). In addition, when using the backpropagation method to obtain the derivative as described above, the required calculation time does not increase significantly with the increase in the number of layers (and thus the increase in structural parameters). Therefore, the micro-nanostructure 100 can be divided into more layers, thereby further improving the accuracy of adjusting the structural parameters.

[0044] After obtaining the derivative of the loss function with respect to the structural parameters of each layer as described above, at block 210, the structural parameters of each of the multiple layers are adjusted using the derivative so that the adjusted micro-nanostructure has a predetermined photon response to incident light. According to the method 200 of an embodiment of the present invention, by utilizing the loss function to adjust the structural parameters of the micro-nanostructure 100 through backpropagation, the accuracy and efficiency of the adjustment can be improved. Thus, a micro-nanostructure having a desired (predetermined) photon response can be obtained.

[0045] The following will refer to Figures 4 to 9 An example of the method 200 for adjusting structural parameters of a micro-nanostructure according to the present invention will be further described. Figure 4 FIG. 4 is a flow chart illustrating a process 400 of determining an actual photon response of a micro-nanostructure to incident light in the method 200 according to an embodiment of the present invention. Figure 4 The process 400 shown corresponds to Figure 2 As shown in box 204. Figure 4 As shown, at block 402, an eigenvector is determined for each of the plurality of layers 300_1 to 300_N of the micro-nanostructure 100 using structural parameters. The elements in the eigenvector represent the eigenmodes of the corresponding layer. Furthermore, each layer has the same number of eigenmodes, for example, equal to the number of stages of the Fourier expansion in the rigorous coupled wave algorithm (e.g., m, where m is an integer greater than 1).

[0046] At block 404, for each two adjacent layers in the plurality of layers, an interlayer transmission matrix for incident light is determined between each two adjacent layers using electromagnetic field boundary conditions. At block 406, an actual photon response is determined using all eigenvectors of the plurality of layers and all interlayer transmission matrices. Figure 4 The process shown may, for example, correspond to the process described above for obtaining the photon response using the rigorous coupled-wave algorithm described above.

[0047] The following will refer to Figures 5 to 7 The schematic diagram shown is used to further describe Figure 4 Examples of the eigenmodes, eigenvector matrices, and inter-layer transfer matrices involved in the process. Figure 5 is a schematic diagram illustrating an eigenmode of each of the plurality of layers 300_1, 300_2, 300_3, ..., or 300_N according to an embodiment of the present invention. Figure 5 As shown, each of the multiple layers 300_1, 300_2, 300_3, ..., or 300_N can transmit m eigenmodes 1, 2, 3, ..., m. The forms of the eigenmodes between different layers can be different. The eigenmodes between the upper surface (the surface that the incident light reaches first) and the lower surface (the surface that the incident light reaches later) within the same layer can also be different. In this case, in some embodiments, the eigenvectors of each layer may include a first eigenvector and a second eigenvector.

[0048] In some embodiments, the elements of the first eigenvector may represent eigenmodes at a first surface (upper surface) within the corresponding layer. In some embodiments, the elements of the second eigenvector may represent eigenmodes at a second surface (lower surface) within the corresponding layer. The first surface and the second surface are facing each other, and incident light sequentially passes through the first surface and the second surface along the transmission direction.

[0049] For example, for any i-th layer (ie, layer 300_i) among the plurality of layers 300_1 to 300_N, i is an integer and 1≤i≤N, the first eigenvector Φ corresponding to the first surface i It can be expressed by the following formula (1), and the corresponding second eigenvector Ψ at the second surface i It can be expressed by the following formula (2):

[0050]

[0051] The first eigenvector Φ i The elements in represent the eigenmodes at the upper surface of the i-th layer (more specifically, the coefficients corresponding to the eigenmodes, for example, weight coefficients), the subscript i represents the i-th layer, and the superscripts 1, 2, ..., m represent the 1st, 2nd, ..., mth eigenmodes. The second eigenvector Ψ i The elements in represent the eigenmodes at the lower surface of the i-th layer, the subscript i represents the i-th layer, and the superscripts 1, 2, ..., m represent the 1st, 2nd, ..., mth eigenmodes.

[0052] like Figure 5 As shown, in each layer 300_1, 300_2, 300_3, ..., or 300_N of the micro-nanostructure 100, the eigenmode can be considered to propagate independently between the upper surface and the lower surface. In this case, the first eigenvector Φ i With the second eigenvector Ψ i There may be a first association, which is represented by the intra-layer transmission matrix of the corresponding layer 300_i. For example, the first eigenvector Φ i With the second eigenvector Ψ i Through the intra-layer transmission matrix P i The first association established can be expressed by the following formula (3):

[0053]

[0054] The above intra-layer transmission matrix P i is a diagonal matrix, for example, the intra-layer transmission matrix P i It can be expressed by the following formula (4):

[0055]

[0056] The intra-layer transmission matrix P in the above formula (4) is i Each element in represents a phase factor.

[0057] In addition, between two adjacent layers (300_1 and 300_2, 300_2 and 300_3, etc.) of the micro-nanostructure 100, the eigenmodes are coupled to each other (indicated by the cross lines in the figure) and propagate according to the electromagnetic field boundary conditions of Maxwell's equations. In this case, in some embodiments, the second eigenvector of the first layer of each two adjacent layers in the plurality of layers may have a second correlation with the first eigenvector of the second layer. The incident light is transmitted along the transmission direction ( Figure 3 The direction shown by the arrow A) passes through the first layer and the second layer in sequence. For example, the first layer is Figure 5 The second layer is layer 300_i-1, where i is greater than 1. The second association can be represented by the corresponding inter-layer transfer matrix. For example, the second eigenvector Ψ of the first layer i-1 and the first eigenvector Φ of the second layer i Through the inter-layer transmission matrix T i The second association established can be expressed by the following formula (5):

[0058]

[0059] In the above equation (5), the subscripts i and i-1 represent the i-th layer (ie, layer 300_i) and the i-1-th layer (ie, layer 300_i-1). The above inter-layer transmission matrix Ti is determined by the electromagnetic field boundary conditions of Maxwell's equations.

[0060] Depend on Figure 5 It can be seen that the transmission mode within and between layers of the micro-nano structure 100 is similar to the transmission mode within and between layers in the neural network model. Figure 6 is a schematic diagram illustrating the similarity between the micro-nanostructure 100 and the neural network model 600 according to an embodiment of the present invention. Figure 7 is another schematic diagram illustrating the similarity between the micro-nanostructure 100 and the neural network model 600 according to an embodiment of the present invention.

[0061] like Figure 6 As shown, the neural network model 600 may have N layers, each layer having m neurons represented by circles ( Figure 6Only four neurons are shown as an example, but not limited thereto). Each layer 300_1, 300_2, 300_3, ..., or 300_N in the micro-nanostructure 100 can correspond to a corresponding layer 600_1, 600_2, 600_3, ..., or 600_N in the neural network model 600. The eigenmode of each layer in the micro-nanostructure 100 can correspond to a corresponding neuron in the corresponding layer of the neural network model 600. In each layer of the neural network model 600, data propagates independently in each neuron. Between adjacent layers of the neural network model 600, the outputs of the upper-layer neurons are mutually coupled and transmitted to the neurons in the lower layer.

[0062] like Figure 7 As shown, for the i-th layer 600_i of the neural network model, the input vector Y at the layer input i It can be expressed by the following formula (6): the output vector X at the layer output i It can be expressed by the following formula (7):

[0063]

[0064] Input vector Y i The elements in represent the inputs of the neurons in the i-th layer 600_i, where the subscript i represents the i-th layer and the superscripts 1, 2, ..., m represent the 1st, 2nd, ..., mth neurons. i The elements in represent the outputs of the neurons in the i-th layer 600_i, the subscript i represents the i-th layer, and the superscripts 1, 2, ..., m represent the 1st, 2nd, ..., mth neurons. i And the output vector X i and the first eigenvector Φ of the i-th layer of the micro-nanostructure 100 i and the second eigenvector Ψ i Have similar forms.

[0065] In addition, the input vector Y i With the output vector X i The activation function matrix F i The following relationship is established:

[0066]

[0067] The above activation function matrix F i It is also a diagonal matrix, which is consistent with the intra-layer transmission matrix P shown in the above equations (3) and (4) i Have similar forms.

[0068] In addition, between two adjacent layers (e.g., layer 600_i-1 and layer 600_i) of the neural network model 600, the output vector X of the upper layer (layer 600_i-1) is i-1 With the input vector Y of the lower layer i There can be a weight matrix W between i The relationship shown in the following formula (9) is established:

[0069]

[0070] In the above formula (9), the subscripts i and i-1 represent the i-th layer (i.e., layer 600_i) and the i-1-th layer (i.e., layer 600_i-1), and i is an integer greater than 1. As can be seen from the above, formulas (6)-(9) have similar forms to formulas (1)-(3) and formula (5), that is, the micro-nano structure 100 and the neural network model 600 have similar transmission methods. Therefore, in the Figure 4 When the process 400 shown obtains the actual photon response of the micro-nanostructure 100, the loss function for the micro-nanostructure 100 can be obtained in a manner similar to that of obtaining the loss function for the neural network model 600. Thereafter, the derivative of the loss function for the micro-nanostructure 100 can be obtained in a manner similar to that of obtaining the derivative of the loss function for the neural network model 600.

[0071] Figure 8 8 is a flowchart illustrating a process 800 of obtaining a derivative of a loss function for a micro-nanostructure 100 with respect to a structural parameter of each layer according to an embodiment of the present invention. Figure 8 The process shown can correspond to Figure 2 208 in the block. Figure 8 As shown, at block 802, a direction of propagation ( Figure 3 In the direction opposite to the direction indicated by the arrow A), according to the chain rule, through the above-mentioned first association and second association, the first derivative of the loss function (hereinafter represented by "L") with respect to the first eigenvector of each layer and the second derivative with respect to the second eigenvector of each layer are determined in sequence.

[0072] For example, as shown in the following equation (10), first for the last layer (layer 300_N), use the first association (intra-layer transmission matrix P N ) Get the loss function L relative to the first eigenvector Φ of the last layer N and the second eigenvector Ψ N The first and second derivatives of :

[0073]

[0074] In the above equation (10), the left matrix represents the first derivative of the last layer (layer 300_N), and the right matrix represents the second derivative of the last layer (layer 300_N). Then, we can use the chain rule to use the second correlation (inter-layer transmission matrix T N ) Get the second eigenvector Ψ of the loss function L relative to the previous layer (layer 300_N-1) of the last layer N-1 The second derivative of is shown in the following formula (11):

[0075]

[0076] In the above equation (11), the left side represents the second derivative of the previous layer (layer 300_N-1) of the last layer. Then, we can use the chain rule to use the first correlation (intra-layer transmission matrix P N-1 ) Get the loss function L relative to the first eigenvector Φ of the previous layer of the last layer N-1 The first derivative of is shown in the following formula (12):

[0077]

[0078] In equation (12), the left side represents the first derivative of the layer above the last layer (layer 300_N-1). Similarly, the first derivative of the loss function L with respect to the first eigenvector of each layer and the second derivative with respect to the second eigenvector of each layer can be determined in the direction opposite to the propagation direction according to the chain rule.

[0079] After obtaining the first and second derivatives for each layer as above, Figure 8 At block 804, for each layer, the first derivative and the second derivative are used to determine the derivative of the loss function with respect to the structural parameters. Here, since the eigenmodes (and therefore the first eigenvector and the second eigenvector) of each layer of the micro-nanostructure 100 are associated with the structural parameters of the layer, after determining the first derivative of the loss with respect to the first eigenvector and the second derivative with respect to the second eigenvector, the derivative of the loss function with respect to the structural parameters of the layer can be determined. For example, it can be determined by any known method.

[0080] After determining the derivative of the loss function with respect to the structural parameters of each layer, we can Figure 2 As shown at block 210 , the determined derivatives are used to adjust structural parameters. Figure 9 is a flow chart illustrating a process 900 for adjusting structural parameters using derivatives according to an embodiment of the present invention. Figure 9 The process 900 shown is Figure 2 An example of method 200 is shown. Figure 9 Boxes 202, 204, 206 and 208 in Figure 2 Boxes 202, 204, 206 and 208 in are the same. Figure 9 Block 210 in may include block 902 , block 904 , block 906 , block 908 , and block 910 .

[0081] At block 902, for each of the plurality of layers 300_1 to 300_N of the micro-nanostructure 100, the structural parameters are adjusted using the derivatives determined at block 208 to obtain a currently adjusted micro-nanostructure. At block 904, the current actual photon response of the currently adjusted micro-nanostructure to the incident light is determined. For example, the above reference Figure 4 The current actual photon response is determined in the manner described. At block 906, it is determined whether the current error of the current actual photon response relative to the predetermined photon response is within a predetermined error range. The predetermined error range can be set differently according to different accuracy requirements.

[0082] In the case where it is determined at block 906 that the current error is within the predetermined error range ( Figure 9 In the case where it is determined at block 906 that the current error is not within the predetermined error range ( Figure 9 If the error is "No" in the error response, a current loss function representing the current error (e.g., the current function value of the loss function as a mean square error function) may be determined at block 910, and the process returns to block 208 to use the current loss function as the loss function to re-adjust the current structural parameters of each of the multiple layers 300_1 to 300_N of the micro-nanostructure 100. The above process is repeated until the adjusted micro-nanostructure has a predetermined photon response to the incident light. In this way, a micro-nanostructure having a predetermined photon response can be efficiently and accurately obtained.

[0083] Figure 10 FIG. 1 is a block diagram of an apparatus 1000 for adjusting structural parameters of a micro-nano structure 100 according to an embodiment of the present invention. Figure 10As shown, the apparatus 1000 may include a first unit 1002, a second unit 1004, a third unit 1006, a fourth unit 1008, and a fifth unit 1010. The first unit 1002 is configured to divide the micro-nanostructure into multiple layers, each of which has corresponding structural parameters. The second unit 1004 is configured to determine the actual photon response of the micro-nanostructure to incident light, where the incident light sequentially passes through the multiple layers along a transmission direction. The third unit 1006 is configured to determine a loss function representing the error of the actual photon response relative to a predetermined photon response. The fourth unit 1008 is configured to obtain, based on the loss function, a derivative of the loss function with respect to the structural parameters of each layer using a backpropagation method with respect to the transmission direction. The fifth unit 1010 is configured to use the derivative to adjust the structural parameters of each layer in the multiple layers so that the adjusted micro-nanostructure has a predetermined photon response to the incident light.

[0084] The above-mentioned apparatus 1000 can execute the above-mentioned reference Figure 2 The method 200 described and referenced Figure 4 、 Figure 8 and Figure 9 The processes 400, 800, and 900 are not described in detail here. The apparatus 1000 according to an embodiment of the present invention utilizes a loss function to adjust the structural parameters of the micro-nanostructure 100 via backpropagation, thereby improving the accuracy and efficiency of the adjustment. Thus, a micro-nanostructure having a desired (predetermined) photon response can be obtained.

[0085] Figure 11 Shown is a schematic block diagram of an example device 1100 suitable for implementing an embodiment of the present disclosure. The controller above can be implemented using device 1100. As shown, device 1100 includes a processor 1101, which can be loaded into a computer program instruction in a random access memory (RAM) 1103 according to a computer program instruction stored in a read-only memory (ROM) 1102, to perform various appropriate actions and processes. In RAM 1103, various programs and data required for memory 1000 operations can also be stored. Processor 1101, ROM 1102, and RAM 1103 are connected to each other via bus 1104. Input / output (I / O) interface 1105 is also connected to bus 1104.

[0086] The various processes and procedures described above, such as method 200 and processes 400, 800, and 900, may be executed by processor 1101. For example, in some embodiments, method 200 and processes 400, 800, and 900 may be implemented as a computer software program tangibly embodied on a machine-readable medium. In some embodiments, part or all of the computer program may be loaded and / or installed onto device 1100 via ROM 1102. When the computer program is loaded into RAM 1103 and executed by processor 1101, one or more actions of method 200 and processes 400, 800, and 900 described above may be performed.

[0087] The various processes and processing described above, such as methods, may be performed at a computing device. The computing device includes, for example: at least one processor (at least one graphics processor and at least one central processing unit); and a memory in communication with the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor. In some embodiments, the method may be implemented as a computer software program or program product, which is tangibly contained in a machine-readable medium. In some embodiments, part or all of the computer program may be loaded and / or installed on a computing device via a read-only memory (ROM) and / or a communication unit. When the computer program is loaded into a random-access memory (RAM) and executed by the GPU and the CPU, one or more actions of the method described above may be performed.

[0088] The present invention may be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions for performing various aspects of the present invention. The computer-readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof.

[0089] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. Various aspects of the present invention are described herein with reference to the flow charts and / or block diagrams of the methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each block of the flow charts and / or block diagrams, and the combination of the blocks in the flow charts and / or block diagrams, can be implemented by computer-readable program instructions.

[0090] These computer-readable program instructions can be provided to a central processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine such that when these instructions are executed by the central processing unit of the computer or other programmable data processing device, a device is generated that implements the functions / actions specified in one or more blocks in the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, where these instructions cause the computer, programmable data processing device, and / or other device to operate in a specific manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0091] The flow charts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the prescribed logical function. In some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented using a dedicated hardware-based system that performs the prescribed function or action, or can be implemented using a combination of dedicated hardware and computer instructions.

[0092] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this application can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this application can be achieved. This is not a limitation herein.

[0093] The above specific embodiments do not constitute a limitation on the scope of protection of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors.

Claims

1. A method for adjusting structural parameters of a micro-nanostructure, comprising: Dividing the micro-nano structure into a plurality of layers, each of the plurality of layers having corresponding structural parameters; determining an actual photon response of the micro-nanostructure to incident light, the incident light sequentially passing through the plurality of layers along a transmission direction; determining a loss function representing an error of the actual photon response relative to a predetermined photon response; According to the loss function, using a back-propagation method relative to the transmission direction, obtaining a derivative of the loss function with respect to the structural parameters of each layer; as well as For each layer of the plurality of layers, the structural parameters are adjusted using the derivatives so that the adjusted micro-nanostructure has the predetermined photon response to the incident light.

2. The method according to claim 1, wherein determining the actual photon response of the micro-nanostructure to incident light comprises: For each layer of the plurality of layers, determining an eigenvector using the structural parameters, elements of the eigenvector representing eigenmodes of the corresponding layer; For every two adjacent layers in the plurality of layers, determining an interlayer transmission matrix for the incident light between every two adjacent layers by using an electromagnetic field boundary condition; as well as The actual photon response is determined by all eigenvectors of the multiple layers and all inter-layer transfer matrices.

3. The method according to claim 2, wherein the eigenvectors include a first eigenvector and a second eigenvector, The elements of the first eigenvector represent the eigenmodes at the first surface inside the corresponding layer, The elements of the second eigenvector represent eigenmodes at a second surface inside the corresponding layer, and The first surface and the second surface are surfaces facing each other, and the incident light passes through the first surface and the second surface in sequence along the transmission direction.

4. The method according to claim 3, wherein the first eigenvector has a first correlation with the second eigenvector, the first correlation being represented by an intra-layer transfer matrix of the corresponding layer, The second eigenvector of the first layer of each two adjacent layers in the plurality of layers has a second correlation with the first eigenvector of the second layer, the second correlation being represented by a corresponding inter-layer transfer matrix, and The incident light passes through the first layer and the second layer in sequence along the transmission direction.

5. The method according to claim 4, wherein obtaining, according to the loss function, a derivative of the loss function with respect to the structural parameter of each layer using a backpropagation method relative to the transmission direction comprises: In a direction opposite to the propagation direction, sequentially determining a first derivative of the loss function with respect to a first eigenvector of each layer and a second derivative with respect to a second eigenvector of each layer through the first association and the second association according to the chain rule; as well as For each layer, the derivative of the loss function with respect to the structural parameter is determined using the first derivative and the second derivative.

6. The method according to claim 1 , wherein for each layer in the plurality of layers, adjusting the structural parameter using the derivative so that the adjusted micro-nanostructure has the predetermined photon response to the incident light comprises: For each layer of the plurality of layers, adjusting the structural parameters using the derivative to obtain a currently adjusted micro-nanostructure; determining a current actual photon response of the currently adjusted micro-nanostructure to the incident light; determining whether a current error of the current actual photon response relative to the predetermined photon response is within a predetermined error range; as well as In response to determining that the current error is within the predetermined error range, it is determined that the currently adjusted micro-nanostructure has the predetermined photon response to the incident light.

7. The method according to claim 6, wherein for each layer of the plurality of layers, adjusting the structural parameters using the derivative so that the adjusted micro-nanostructure has the predetermined photon response to the incident light further comprises: In response to determining that the current error is not within the predetermined error range, determining a current loss function representing the current error; as well as The current loss function is used as the loss function to re-adjust the current structural parameters of each layer in the plurality of layers until the adjusted micro-nanostructure has the predetermined photon response to the incident light.

8. The method of any one of claims 1 to 7, wherein the photon response comprises at least one of: reflectivity, transmittance, phase shift, and polarization. 9 . The method according to claim 1 , wherein the plurality of layers are virtually divided layers, and the micro-nano structure is an integrated structure.

10. The method according to any one of claims 1 to 7, wherein the micro-nanostructure is used for at least one of the following: a grating, an optical film, a metasurface, a reflector, a polarizer, and a sensor. 11 . The method according to claim 1 , wherein the structural parameter of each layer of the plurality of layers comprises at least one of the following: thickness, vertex coordinates of each vertex, and length of each edge.

12. The method of any one of claims 1 to 7, wherein the loss function represents a mean square error between the actual photon response and the predetermined photon response.

13. A device for adjusting structural parameters of a micro-nanostructure, comprising: A first unit is configured to divide the micro-nano structure into a plurality of layers, each of the plurality of layers having corresponding structural parameters; a second unit configured to determine an actual photon response of the micro-nanostructure to incident light, the incident light sequentially passing through the plurality of layers along a transmission direction; a third unit configured to determine a loss function representing an error of the actual photon response relative to a predetermined photon response; A fourth unit is configured to obtain, according to the loss function, a derivative of the loss function with respect to the structural parameter of each layer using a backpropagation method relative to the transmission direction; as well as A fifth unit is configured to adjust the structural parameter of each layer of the plurality of layers using the derivative so that the adjusted micro-nanostructure has the predetermined photon response to the incident light.

14. A computer program product comprising a computer program, wherein when the computer program is executed by a machine, the method according to any one of claims 1 to 12 is performed.

Images