Thermal electron detector design method and system based on lightweight gradient elevator

Abstract

The invention discloses a thermal electron detector design method and system based on a lightweight gradient elevator, and relates to the technical field of device design, and the method is characterized in that a lightweight gradient elevator model is constructed according to initial sample data training of a layered nanometer photonic device, and each sample data comprises geometric parameters and a corresponding reflection spectrum; and inputting the target reflection spectrum into a pre-constructed lightweight gradient elevator model to obtain predicted geometric parameters. According to the method, the lightweight gradient elevator is used for fitting the mapping relation between the structural parameters and the optical response of the hot electron detector, and then reverse design is carried out according to needs; according to the method, the function relationship between the structural parameters and the optical response of the Tamm-state plasma thermal electron detector can be learned, and the working performance of the device is optimized through prediction and on-demand design; the lightweight gradient elevator can pre-process the data through an own feature importance algorithm, and selects the data with effective contribution.

Description

Technical Field

[0001] The present invention relates to the technical field of device design, and more specifically, to a thermal electron detector design method and system based on a lightweight gradient boosting machine. Background Art

[0002] As one of the important branches of machine learning, deep learning has changed the development direction of many fields and industries. The design problem of neural networks is usually transformed into a neural network that establishes a functional relationship between input and output. However, a simple neural network cannot converge when faced with samples of one-to-many mapping relationships. Nanophotonic devices are a type of device with broad application prospects, and their performance and characteristics are affected by their structural parameters. Among them, thermal electron detectors provide research ideas for optical communications with their low development cost, high sensitivity, and easy integration.

[0003] In traditional design methods, finite element analysis is usually used to verify the permutations and combinations of data. However, this method has the problem of occupying computing resources and consuming a lot of time, and it is impossible to traverse all possible parameter combinations. In recent years, the use of deep learning algorithms has made considerable progress in the auxiliary design of nanophotonic devices, which has greatly improved the design efficiency. However, the current deep learning design is a many-to-many prediction method, and the objective function obtained is to obtain the fitting function based on the overall target, which makes it difficult to meet the accuracy of data in some areas. In addition, the current application of deep learning to data is generally direct, without preprocessing the original data, and it is impossible to extract effective features, which may lead to a decrease in overall accuracy.

[0004] Therefore, how to study and design a method for designing optical devices for Tamm state plasma thermal electron detectors that can overcome the above-mentioned defects is a problem that we urgently need to solve. Summary of the invention

[0005] In order to address the deficiencies in the prior art, the purpose of the present invention is to provide a thermal electron detector design method and system based on a lightweight gradient boosting machine, a method that uses a lightweight gradient boosting machine to fit the mapping relationship between the structural parameters and optical responses of the thermal electron detector and then performs reverse design on demand; this method can learn the functional relationship between the structural parameters and optical responses of the Tamm state plasma thermal electron detector, and optimize the working performance of the device through prediction and on-demand design; the lightweight gradient boosting machine can pre-process the data through its own feature importance algorithm and select data with effective contribution.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions:

[0007] In a first aspect, a method for designing a thermal electron detector based on a lightweight gradient lifting machine is provided, comprising the following steps:

[0008] A lightweight gradient boosting machine model is constructed based on the initial sample data training of layered nanophotonic devices, where each sample data includes geometric parameters and corresponding reflection spectra;

[0009] The target reflectance spectrum is input into a pre-built lightweight gradient boosting machine model to obtain the predicted geometric parameters.

[0010] Furthermore, the lightweight gradient boosting machine model includes:

[0011] The forward prediction network is trained and constructed by taking the geometric parameters in the initial sample data as input and the reflectance spectrum in the initial sample data as output;

[0012] The reverse design network is trained and constructed by taking the reflection spectrum in the valid sample data as input and the geometric parameters in the valid sample data as output;

[0013] The valid sample data are as follows:

[0014] The contribution value of the eigenvalue data in the initial sample data is calculated by the feature importance method, and the valid sample data is screened out from the initial sample data by the contribution value.

[0015] Furthermore, the process of obtaining the valid sample data is specifically as follows:

[0016] The contribution values ​​of the reflectivity points corresponding to the reflectivity spectrum and geometric parameters are calculated by the feature importance method of the lightweight gradient boosting machine;

[0017] If the contribution value of a reflectivity point is greater than 1, the reflection spectrum and geometric parameters corresponding to the corresponding reflectivity point are valid data, and all valid data constitute valid sample data.

[0018] Furthermore, the process of obtaining the initial sample data is specifically as follows:

[0019] The nanostructure of layered Tamm-state plasma is constructed using the time-domain finite element difference method to obtain the geometric parameters.

[0020] And, by characterizing the optical response of the nanostructure, the corresponding reflection spectrum is obtained.

[0021] Furthermore, the lightweight gradient boosting machine model gradually reduces the learning rate during the training process, and increases the depth of the tree or adjusts the number of leaf nodes.

[0022] Furthermore, the lightweight gradient boosting machine model uses a many-to-one cyclic prediction method to achieve cyclic prediction of each geometric parameter;

[0023] Among them, each geometric parameter corresponds to an objective function.

[0024] Furthermore, the predicted geometric parameters include the number of distributed Bragg reflector layers and the thicknesses of the corresponding layers of gold, titanium dioxide, and silicon dioxide.

[0025] Secondly, a thermal electron detector design system based on a lightweight gradient lifting machine is provided, including:

[0026] A model training module is used to train and construct a lightweight gradient boosting machine model based on initial sample data of layered nanophotonic devices, each of which includes geometric parameters and corresponding reflection spectra;

[0027] The parameter prediction module is used to input the target reflectance spectrum into the pre-built lightweight gradient boosting machine model to obtain the predicted geometric parameters.

[0028] In a third aspect, a computer terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a thermal electron detector design method based on a lightweight gradient boosting machine as described in any one of the first aspects is implemented.

[0029] In a fourth aspect, a computer-readable medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement a thermal electron detector design method based on a lightweight gradient boosting machine as described in any one of the first aspects.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. The present invention provides a method for designing a thermal electron detector based on a lightweight gradient boosting machine, which uses a lightweight gradient boosting machine to fit the mapping relationship between the structural parameters and optical response of the thermal electron detector and then performs reverse design on demand; the method can learn the functional relationship between the structural parameters and optical response of the Tamm state plasma thermal electron detector, and optimize the working performance of the device through prediction and on-demand design; the lightweight gradient boosting machine can pre-process the data through its own feature importance algorithm and select data with effective contribution;

[0032] 2. The present invention realizes a many-to-one prediction method and establishes an objective function for each predicted eigenvalue, specifically solving the problem of computing resource consumption in traditional design. This method can also explore more possible combinations of structural parameters, which improves the efficiency of accurately designing thermal electron detectors on demand. This method provides ideas for the research of Tamm state plasma thermal electron detectors. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0034] Figure 1 is an exemplary schematic diagram of a Tamm-state plasma hot electron detector in Example 1 of the present invention;

[0035] Figure 2 A side view and collective parameters of a sensor for detecting polarized light in Example 1 of the present invention;

[0036] Figure 3 The design architecture and flow chart of the lightweight gradient boosting machine in Example 1 of the present invention;

[0037] Figure 4 This is a flow chart of effective processing of data features of a lightweight gradient boosting machine in Example 1 of the present invention;

[0038] Figure 5 Schematic diagram of the importance score of data features of reflection points in Example 1 of the present invention;

[0039] Figure 6 This is a flow chart of many-to-one cyclic prediction in Embodiment 1 of the present invention;

[0040] Figure 7 This is a diagram showing the effect of constructing parameter mapping of nanophotonic devices using a forward prediction network in Example 1 of the present invention;

[0041] Figure 8 This is a diagram showing the effect of the inverse design network constructing the parameter mapping of the nanophotonic device in Example 1 of the present invention;

[0042] Fig. 9 This is a diagram showing the effect of designing a thermal electron detector on demand in reverse design in Example 1 of the present invention;

[0043] Fig.10 A detailed flow chart of lightweight on-demand design and verification in Embodiment 1 of the present invention;

[0044] Fig.11 Detailed information on the thermal electron detector designed on demand in Example 1 of the present invention;

[0045] Fig.12 This is a system block diagram of Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0046] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0047] Embodiment 1: A method for designing a thermal electron detector based on a lightweight gradient lifting machine comprises the following steps:

[0048] Step 1: A lightweight gradient boosting machine model is constructed based on the initial sample data of layered nanophotonic devices. Each sample data includes geometric parameters and corresponding reflection spectra.

[0049] Step 2: Input the target reflectance spectrum into the pre-built lightweight gradient boosting machine model to obtain the predicted geometric parameters.

[0050] In step 1, the lightweight gradient boosting machine model includes a forward prediction network and a reverse design network. The forward prediction network is trained and constructed with the geometric parameters in the initial sample data as input and the reflectance spectrum in the initial sample data as output; the reverse design network is trained and constructed with the reflectance spectrum in the effective sample data as input and the geometric parameters in the effective sample data as output.

[0051] Specifically, (1) the optical response of 4- and 5-layer layered Tamm-state plasmas is characterized by using the time-domain finite element difference method, and the scanning parameters are the thickness of the gold layer (90-100nm, 9 times), silicon dioxide (60-70nm, 11 times), titanium dioxide layer (70-80nm, 11 times) and the number of layers (4, 5). The required initial sample data is 2178, and the reflection spectrum consists of 253 reflection points. The data set is divided into 80% training set and 20% test set. The forward prediction network is trained using the training set data. The mapping relationship between the geometric parameters and the optical response for the design is obtained.

[0052] (2) The reverse design network is trained using the divided training set data. The reverse design network includes input: optical response; decision tree; parallel optimization; output predicted geometric parameters. The optical response input to the reverse design network is traversed through the data one after another, and the statistics are accumulated in the histogram according to the discretized value as the index. After traversing the data once, the histogram accumulates the required statistics, and then the optimal segmentation point is traversed according to the discrete value of the histogram.

[0053] In data preprocessing, the contribution values ​​of the reflectance spectrum (2178×253) and geometric parameters (2178×4) are calculated by the feature importance method of the lightweight gradient boosting machine. The contribution value greater than 1.0 is defined as valid data. The valid data set is the mapping relationship between (2178×201×4) and the trained data, and the predicted value is finally output.

[0054] (3) In the on-demand design stage, the predicted geometric parameters are calculated by the finite element difference method to obtain the optical response data. This step is to verify the predictive performance of the model. By comparing the calculated optical response with the target optical response designed on demand, the design performance of the model is determined to provide a basis for the actual production of the sample. A hot electron detector with an absorbance greater than 70% (reflectivity less than 30%) at 660nm is designed. The reflection spectrum of 201×1 is input into the lightweight gradient lifting machine model of the reverse design. The output geometric parameters are 100nm for gold, 77nm for titanium dioxide, 60nm for silicon dioxide, and 5 layers for the distributed Bragg reflector. The reflection spectrum data of the parameter is calculated by finite element difference and compared with the on-demand design reflection diagram. The absorption peak (reflection) is at 660nm. This shows that the model can explore the target optical response as much as possible to obtain the design geometric parameters, thereby realizing the reverse design of the Tamm state plasma hot electron detector.

[0055] For example, the device designed by the present invention selects silicon wafer as the research substrate and metal gold as the main material of the absorber. The metal has a high melting point and stability, and has good plasmon performance. Based on thermal plasmons, it guarantees zero transmittance. The distributed Bragg reflector is composed of silicon dioxide and titanium dioxide. The gold backplane and the distributed Bragg reflector can form Tamm state plasma. Figure 1 As shown, the Tamm-state plasma hot electron detector provided by the present invention includes a substrate layer, a gold layer, and a TiO2 / SiO2 alternating layer. The light source is required to be a plane wave, with a vertical incident unit structure from the top, and the wavelength range of the light source is 600nm-1000nm, and the distance between the light source and the device is greater than one wavelength. The boundary conditions of the simulation interval are periodic boundary conditions in the x-direction and the y-direction, and the z-direction is a perfect matching layer. A reflectivity monitor is set on the light source to obtain the reflectivity distribution between 500nm-1000nm, and mapped to 201 point outputs. The thickness of each layer is between 100nm-300nm, and the other areas are reserved as air. Since the substrate is a silicon wafer, the transmittance is almost 0, and there is no need to set a transmittance monitor, so the absorbance value A=1-RT, T=0, such as Figure 2 As shown, its side view, including parameters and materials.

[0056] After the required simulation conditions are set, the geometric parameter range is composed of an initial data set or arranged and combined in a regular pattern to ensure a reasonable coverage range. The specific implementation scheme of the present invention has 11×11×9=1089 label values ​​of the initial sample data, and the reflection spectrum corresponding to its geometric parameters is obtained by numerical simulation through the finite element difference method. The initial sample data consists of geometric parameters and corresponding reflection spectra. Four more layers are taken for value selection, and numerical simulation is also performed through the time-domain finite element difference method to obtain a data set consisting of 1089 groups of structural parameters and optical responses as a test and verification data set, and the total data set is 2178. In order to enable the model to better learn the mapping relationship between the input data X and Y, the label values ​​need to be normalized. This specific implementation scheme adopts linear normalization, that is:

[0057] like Figure 3 As shown in the figure, during the training process, the lightweight gradient boosting machine model calculates the loss value in each iteration based on the difference between the current model prediction value and the true label. Specifically, the model uses the mean square error as the loss function to measure the prediction error of the current model. The process calculates the loss for each sample in each training iteration and updates the model parameters through the gradient boosting algorithm.

[0058] In multi-label regression problems, if the model outputs multiple sets of prediction parameters, each set of predictions will be compared with the corresponding label value to calculate the loss value. Assuming there are multiple labels, the loss value will be calculated separately for each set of predictions, and the minimum loss value will be taken as the current optimal loss value for update, which helps the model avoid error accumulation due to the complexity of the labels. Set the loss value and determine the minimum loss value by calculating the mean square error between each set of structural parameters and the label. The purpose of this step is to ensure that when faced with multiple labels, the model can match each label one by one and minimize the overall error.

[0059] In order to better obtain the optimal data, the lightweight gradient boosting machine model adopts Figure 4 The feature importance method measures the contribution of each absorption point feature in all decision trees. If the feature value score is greater than 1.0, it means that the feature is contributing, which can effectively extract data in forward prediction and provide input data for reverse design. In actual testing, Figure 5 shown. Figure 5 (A) shows the contribution values ​​of 253 reflection points. Figure 5 (B) is sorted from high to low. As a result, 52 reflection points with importance lower than 1.0 were screened out.

[0060] like Figure 6As shown in the figure, in order to solve the problem of large prediction error in the traditional design of the many-to-many model, the lightweight gradient boosting machine is optimized. The optimized model adopts a many-to-one cyclic prediction method to achieve cyclic prediction of each parameter and construct multiple objective functions. This can avoid errors in some areas caused by overall prediction and improve the accuracy of the prediction.

[0061] During the training process, in order to accelerate the convergence of the model and avoid overfitting, the lightweight gradient boosting machine model dynamically adjusts the learning rate and regularization term. In the initial training, a larger learning rate (such as 0.1) can be set to quickly explore the best solution. As the training progresses, the learning rate is gradually reduced, and the depth of the tree is increased or the number of leaf nodes is adjusted to further improve the accuracy of the model. In the end, the model should reach a smaller loss value and show high prediction accuracy on the test set. Parameter adjustment during training, during the training process, some key hyperparameters of the model need to be adjusted: Learning rate: The learning rate controls the step size of the model parameter update at each iteration. It can be set to 0.1 initially and then adjusted gradually. Number of trees: The complexity of the model can be controlled by setting the number of iterations or the number of trees. It can be set to 100 initially and then adjusted according to the change in loss value. Maximum depth of the tree: The depth of the tree controls the complexity of each tree. It can be set to 6 initially and adjusted according to the performance during training. Number of leaf nodes: The number of leaf nodes affects the capacity of the model. Increasing the number of leaf nodes can improve the expressiveness of the model, but it will also increase the risk of overfitting. Regularization parameters: These parameters control the regularization effect of the model to avoid overfitting. By adjusting these hyperparameters and combining the feedback of the loss value during training, the model can converge to a better prediction result. During training, the model continuously adjusts the structure of the decision tree to minimize the error between the output prediction result and the true label. Finally, the model will make predictions based on the learned patterns and output the final prediction results on the test set or actual data. These prediction results can be further used for various tasks such as optical response and structural parameter prediction.

[0062] After training the forward prediction network, Figure 7 As shown, the prediction effect of the forward prediction network of the present invention is demonstrated, ensuring the accuracy and high generalization ability of the forward prediction network. The forward prediction network is used to train the reverse design network according to the algorithm block diagram, and the network is designed according to the requirements of this specific implementation example. Figure 8 As shown in FIG. , it is the fitting coefficient of each reflection point of the absorption spectrum, and the fitting coefficient is greater than 0.95. In this specific implementation example, Fig. 9 Four sets of geometric parameters output by the inverse design network are shown. It can be seen that the error between the predicted parameters and the actual parameters is small (accuracy: 0.99). Fig.10The flowchart of the on-demand device design is shown. For example: you need to design a hot electron detector with 660nm absorption, and input the corresponding optical response spectrum into the lightweight gradient boosting machine. At this time, the model is after the previous training. Get the output geometric parameters of the model and prepare the device according to the obtained geometric parameters. Before preparing the device, the optical response can be calculated by the finite element difference method, and the on-demand optical response is compared. The geometric parameter values ​​designed on demand are that the thickness of gold is 100nm, the thickness of titanium dioxide is 77nm, the thickness of silicon dioxide is 60nm, and the number of layers of the designed distributed Bragg reflector is 5. Verify the accuracy of the model design, calculate the obtained geometric parameters as the finite time-domain difference method, and compare the calculated results with the input. The absorption spectrum is calculated as A=1-RT, T is 0.

[0063] like Fig.11 As shown, it includes input reflection spectrum data, predicted geometric parameters and simulated absorption spectrum. It is not difficult to see that the lightweight gradient lifting machine is conducive to the auxiliary design of Tamm state plasma hot electron detectors, and can realize on-demand design according to the required target optical response, which has certain inspirational significance. In today's rapid development of nanophotonic devices, the demand for high-performance device design is constantly strong. The design method of the present invention has a strong support power and reference value for the future extensive development of nanophotonic device design.

[0064] As one of the machine learning methods, ensemble learning improves the overall prediction performance by combining the prediction results of multiple models, and builds a more powerful ensemble model by combining multiple weak learners. The advantage of ensemble learning is that it can reduce overfitting, improve the generalization ability of the model, and usually has better performance than a single model. In addition, ensemble learning can also use parallel computing to accelerate the model training process. Compared with neural networks, lightweight gradient boosting machines can only achieve many-to-one predictions, which is in line with the design of Tamm state plasma hot electron detectors, because its absorption and reflection peaks are single, and the geometric parameters are also single and independent. In addition, lightweight gradient boosting machines have the characteristics of fast training and not easy to overfit, which can effectively solve the problems in traditional designs.

[0065] Embodiment 2: A thermal electron detector design system based on a lightweight gradient lifting machine, the system is used to implement the thermal electron detector design method based on a lightweight gradient lifting machine as described in Embodiment 1, such as Fig.12 As shown, it includes a model training module and a parameter prediction module.

[0066] Among them, the model training module is used to train and construct a lightweight gradient boosting machine model based on the initial sample data of layered nanophotonic devices, and each sample data includes geometric parameters and corresponding reflection spectra; the parameter prediction module is used to input the target reflection spectrum into the pre-constructed lightweight gradient boosting machine model to obtain predicted geometric parameters.

[0067] The present invention also records a computer terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the thermal electron detector design method based on a lightweight gradient boosting machine as described in Example 1 is implemented.

[0068] The present invention also records a computer-readable medium on which a computer program is stored. The computer program is executed by a processor to implement the thermal electron detector design method based on a lightweight gradient lifting machine as described in Example 1.

[0069] Working principle: The present invention uses a lightweight gradient boosting machine to fit the mapping relationship between the structural parameters and optical response of the thermal electron detector and then reversely design it on demand; this method can learn the functional relationship between the structural parameters and optical response of the Tamm state plasma thermal electron detector, and optimize the working performance of the device through prediction and on-demand design; the lightweight gradient boosting machine can pre-process the data through its own feature importance algorithm and select data with effective contribution;

[0070] In addition, the present invention realizes a many-to-one prediction method and establishes an objective function for each predicted eigenvalue, which specifically solves the problem of computing resource consumption in traditional design. This method can also explore more possible combinations of structural parameters, which improves the efficiency of accurately designing thermal electron detectors on demand. This method provides ideas for the research of Tamm state plasma thermal electron detectors.

[0071] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0072] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A thermal electron detector design method based on a lightweight gradient boosting machine, characterized in that: The following steps are involved: A lightweight gradient boosting machine model is constructed based on the initial sample data training of layered nanophotonic devices, each of which includes geometric parameters and corresponding reflection spectra; The target reflectance spectrum is input into a pre-built lightweight gradient boosting machine model to obtain the predicted geometric parameters.

2. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 1, characterized in that: The lightweight gradient boosting machine model includes: The forward prediction network is trained and constructed by taking the geometric parameters in the initial sample data as input and the reflectance spectrum in the initial sample data as output; The reverse design network is trained and constructed by taking the reflection spectrum in the valid sample data as input and the geometric parameters in the valid sample data as output; The valid sample data are as follows: The contribution value of the eigenvalue data in the initial sample data is calculated by the feature importance method, and the valid sample data is screened out from the initial sample data by the contribution value.

3. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 2, characterized in that: The specific process of obtaining the valid sample data is as follows: The contribution values ​​of the reflectivity points corresponding to the reflectivity spectrum and geometric parameters are calculated by the feature importance method of the lightweight gradient boosting machine; If the contribution value of a reflectivity point is greater than 1, the reflection spectrum and geometric parameters corresponding to the corresponding reflectivity point are valid data, and all valid data constitute valid sample data.

4. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 1, characterized in that: The process of obtaining the initial sample data is specifically as follows: The nanostructure of layered Tamm-state plasma is constructed using the time-domain finite element difference method to obtain the geometric parameters. And, by characterizing the optical response of the nanostructure, the corresponding reflection spectrum is obtained.

5. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 1, characterized in that: The lightweight gradient boosting machine model gradually reduces the learning rate during the training process, and increases the depth of the tree or adjusts the number of leaf nodes.

6. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 1, characterized in that: The lightweight gradient boosting machine model uses a many-to-one cyclic prediction method to achieve cyclic prediction of each geometric parameter; Among them, each geometric parameter corresponds to an objective function.

7. The method for designing a thermal electron detector based on a lightweight gradient boosting machine according to claim 1, characterized in that: The predicted geometric parameters include the number of distributed Bragg reflector layers and the thickness of the corresponding layers of gold, titanium dioxide, and silicon dioxide.

8. A thermal electron detector design system based on a lightweight gradient boosting machine, characterized by: include: A model training module is used to train and construct a lightweight gradient boosting machine model based on initial sample data of layered nanophotonic devices, each of which includes geometric parameters and corresponding reflection spectra; The parameter prediction module is used to input the target reflectance spectrum into the pre-built lightweight gradient boosting machine model to obtain the predicted geometric parameters.

9. A computer terminal comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the thermal electron detector design method based on the lightweight gradient boosting machine as described in any one of claims 1 to 7 is implemented.

10. A computer readable medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the thermal electron detector design method based on a lightweight gradient boosting machine as described in any one of claims 1 to 7.

Images