Electrical parameter prediction method using convolutional neural network system and related device

Abstract

The invention discloses an electrical parameter prediction method using a convolutional neural network system and related equipment, and the method comprises the steps: obtaining an input process parameter through an input layer; extracting features of the input process parameters via the intermediate layer; and outputting an electrical parameter prediction result for the input process parameters through the output layer. According to the technical scheme provided by the embodiment of the invention, the overfitting problem of a full-connection neural network system can be avoided, so that the accuracy of electrical parameter prediction can be improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of semiconductor manufacturing, and in particular to an electrical parameter prediction method using a convolutional neural network system and related equipment. Background Art

[0002] In recent years, with the rapid development of deep neural networks, both academia and industry have witnessed significant breakthroughs in deep learning across numerous fields. Rapid advancements in fields such as information science, energy, and national defense have placed diverse demands on semiconductor devices. Based on currently mature processes, neural networks offer an option for rapidly predicting device structures or properties, while ensuring accuracy while reducing computational costs and shortening development cycles.

[0003] Currently, the solution of using a fully connected neural network system to predict the electrical parameters of semiconductor devices using process parameters as input has the problem of poor prediction accuracy. Summary of the Invention

[0004] The problem solved by the embodiments of the present invention is to provide an electrical parameter prediction method and related equipment using a convolutional neural network system, which can avoid the overfitting problem of a fully connected neural network system, thereby improving the accuracy of electrical parameter prediction.

[0005] To solve the above problems, an embodiment of the present invention provides an electrical parameter prediction method using a convolutional neural network system. The convolutional neural network system includes an input layer, an intermediate layer, and an output layer. The electrical parameter prediction method includes:

[0006] obtaining the input process parameters via the input layer;

[0007] extracting features of the input process parameters via the intermediate layer;

[0008] The electrical parameter prediction result for the input process parameter is outputted via the output layer.

[0009] Optionally, the intermediate layer includes a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer and a fully connected layer, the first convolutional layer includes M first filters, the second convolutional layer includes M second filters, the third convolutional layer includes M third filters, and the fourth convolutional layer includes M fourth filters;

[0010] The extracting of the features of the input process parameters through the intermediate layer includes: using the M first filters in the first convolutional layer to respectively receive the input electrical parameters, and performing a first convolution operation on the input electrical parameters to obtain first feature data of M channels; using the M second filters in the second convolutional layer to respectively receive the first feature data of the M channels, and performing a second convolution operation on the first feature data of the M channels to obtain second feature data of the M channels; using the M third filters in the third convolutional layer to respectively receive the second feature data of the M channels, and performing a third convolution operation on the second feature data of the M channels to obtain third feature data of the M channels; using the M fourth filters in the fourth convolutional layer to respectively receive the third feature data of the M channels, and performing a fourth convolution operation on the third feature data of the M channels to obtain fourth feature data of the M channels; using the fully connected layer to receive the fourth feature data of the M channels, and performing a fully connected operation on the fourth feature data of the M channels to obtain electrical parameter prediction results for the input process parameters.

[0011] Optionally, the convolution kernel size of the first filter, the convolution kernel size of the second filter, the convolution kernel size of the third filter and the convolution kernel size of the fourth filter are respectively 1*3.

[0012] Optionally, the intermediate layer further includes an activation function layer located between two adjacent convolutional layers in the first convolutional layer, the second convolutional layer, the third convolutional layer, and the fourth convolutional layer;

[0013] The extracting the features of the input process parameters via the intermediate layer also includes: using the activation function layer to perform nonlinear conversion operations on the first feature data of M channels, the second feature data of M channels, and the third feature data of M channels respectively.

[0014] Optionally, the intermediate layer further includes at least one of a pooling layer and a batch normalization layer located after the activation function layer;

[0015] The extracting of the features of the input process parameters via the intermediate layer also includes: using the pooling layer to pool the multiple transformation features output by the activation function layer to obtain pooled input features; and using the batch normalization layer to perform batch normalization on the pooled input features.

[0016] Accordingly, an embodiment of the present invention further provides an electrical parameter prediction device using a convolutional neural network system, wherein the convolutional neural network system includes an input layer, an intermediate layer, and an output layer, and the electrical parameter prediction device includes:

[0017] a process parameter input unit, configured to obtain the input process parameters via the input layer;

[0018] a parameter feature extraction unit, configured to extract features of the input process parameters via the intermediate layer;

[0019] The processing result output unit is used to output the electrical parameter prediction result for the input process parameter via the output layer.

[0020] Accordingly, an embodiment of the present invention also provides a computer device, characterized in that it includes at least one memory and at least one processor, the memory storing one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of the above items.

[0021] Accordingly, an embodiment of the present invention also provides a computer program product, including a computer program / instruction, which, when executed by a processor, is used to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of the above.

[0022] Correspondingly, an embodiment of the present invention also provides a storage medium, which stores one or more computer instructions, and the one or more computer instructions are used to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of the above items.

[0023] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0024] An embodiment of the present invention provides an electrical parameter prediction method using a convolutional neural network system, wherein the convolutional neural network system includes an input layer, an intermediate layer, and an output layer. The electrical parameter prediction method includes: obtaining the input process parameters via the input layer; extracting the features of the input process parameters via the intermediate layer; and outputting the electrical parameter prediction results for the input process parameters via the output layer.

[0025] An embodiment of the present invention provides an electrical parameter prediction method using a convolutional neural network system, which performs electrical parameter prediction processing on input process parameters through a preset convolutional neural network system to obtain electrical parameter prediction results for the input process parameters. The convolutional neural network system can avoid overfitting problems, improve the prediction accuracy of electrical parameters, and thus help improve product yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 1 is a flow chart of an embodiment of a method for predicting electrical parameters using a convolutional neural network system provided by the technical solution of the present invention;

[0027] Figure 2 This is a structural diagram of an embodiment of a convolutional neural network system provided by the technical solution of the present invention;

[0028] Figure 3 It is a structural diagram of the middle layer of the convolutional neural network system in the electrical parameter prediction method using the convolutional neural network system provided by the technical solution of the present invention;

[0029] Figure 4 1 is a schematic structural diagram of an embodiment of an electrical parameter prediction device using a convolutional neural network system provided by the technical solution of the present invention;

[0030] Figure 5 This is a schematic diagram of an optional hardware structure of a computer device provided by the technical solution of the present invention. DETAILED DESCRIPTION

[0031] As can be seen from the background art, the solution of using a fully connected neural network system to predict the electrical parameters of semiconductor devices using process parameters as input has the problem of poor prediction accuracy.

[0032] In order to solve the above technical problems, an embodiment of the present invention provides an electrical parameter prediction method using a convolutional neural network system, wherein the convolutional neural network system includes an input layer, an intermediate layer and an output layer, and the electrical parameter prediction method includes: obtaining the input process parameters via the input layer; extracting the features of the input process parameters via the intermediate layer; and outputting the electrical parameter prediction results for the input process parameters via the output layer.

[0033] The electrical parameter prediction method using a convolutional neural network system provided in an embodiment of the present invention performs electrical parameter prediction processing on the input process parameters through a preset convolutional neural network system to obtain electrical parameter prediction results for the input process parameters. Since the convolutional neural network system can avoid the overfitting problem, it can improve the prediction accuracy of the electrical parameters, which is beneficial to improving product yield.

[0034] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0035] Figure 1 A flow chart of an embodiment of a method for predicting electrical parameters using a convolutional neural network (CNN) system provided by the technical solution of the present invention is shown; Figure 2 The structure diagram of an embodiment of the convolutional neural network system provided by the technical solution of the present invention is shown. Figure 1 and Figure 2A method for predicting electrical parameters using a convolutional neural network system may include the following steps:

[0036] Step S110: obtaining the input process parameters via the input layer;

[0037] Step S120: extracting features of the input process parameters via the intermediate layer;

[0038] Step S130: Outputting electrical parameter prediction results for the input process parameters via the output layer.

[0039] Please continue to refer to Figure 1 and Figure 2 , executing step S110, obtaining the input process parameters via the input layer.

[0040] The input process parameters are obtained through the input layer 210 , providing a basis for subsequently extracting features of the input process parameters through the intermediate layer 220 .

[0041] The input process parameters refer to parameters related to the manufacturing process of the semiconductor device, which can be set by the designer of the semiconductor device according to actual needs and are not limited here.

[0042] In an exemplary embodiment, the input process parameters are process parameters of a complementary metal oxide semiconductor (CMOS) device. The parameters related to the manufacturing process of the CMOS device may include one or more of process parameters of a gate oxide layer formation process, process parameters of an N-type polysilicon gate pre-doping process, process parameters of a polysilicon gate etching process, process parameters of an offset spacer formation process, process parameters of a pocket implantation process, process parameters of a spacer formation process, process parameters of a source / drain implantation process, and process parameters of an annealing process.

[0043] Please continue to refer to Figure 1 and Figure 2 , executing step S120, extracting the features of the input process parameters through the intermediate layer.

[0044] The features of the input process parameters are extracted via the intermediate layer 220 , providing a basis for subsequently outputting electrical parameter prediction results for the input process parameters via the output layer 230 .

[0045] In an exemplary embodiment, referring to Figure 3The intermediate layer 220 described includes a first convolutional layer 221, a second convolutional layer 222, a third convolutional layer 223, a fourth convolutional layer 224, and a fully connected layer 225. The first convolutional layer 221 includes M first filters (not labeled), the second convolutional layer 222 includes M second filters (not labeled), the third convolutional layer 223 includes M third filters (not labeled), and the fourth convolutional layer 224 includes M fourth filters (not labeled).

[0046] Correspondingly, the step of extracting the features of the input process parameters through the intermediate layer includes: using the M first filters in the first convolution layer 221 to respectively receive the input electrical parameters, and performing a first convolution operation on the input electrical parameters to obtain first feature data of M channels; using the M second filters in the second convolution layer 222 to respectively receive the first feature data of the M channels, and performing a second convolution operation on the first feature data of the M channels to obtain second feature data of M channels; using the M third filters in the third convolution layer 223 to respectively receive the first feature data of the M channels, and performing a second convolution operation on the first feature data of the M channels to obtain second feature data of M channels. The second feature data of the M channels are received, and a third convolution operation is performed on the second feature data of the M channels respectively to obtain the third feature data of the M channels; the M fourth filters in the fourth convolutional layer 224 are used to receive the third feature data of the M channels respectively, and a fourth convolution operation is performed on the third feature data of the M channels respectively to obtain the fourth feature data of the M channels; the fully connected layer 225 is used to receive the fourth feature data of the M channels, and a fully connected operation is performed on the fourth feature data of the M channels to obtain the electrical parameter prediction results for the input process parameters.

[0047] In the first convolutional layer 221, the second convolutional layer 222, the third convolutional layer 223, and the fourth convolutional layer 224, the features extracted by the earlier convolutional layers are relatively low-level. As the convolutional layers deepen, the receptive field of the convolution kernel gradually increases, and it can capture more complex data features. This enables the later convolutional layers to iteratively extract higher-level and more abstract features from the low-level features extracted by the earlier convolutional layers. This hierarchical data feature extraction method is consistent with the human cognitive process and helps improve the accuracy of feature extraction.

[0048] Furthermore, the convolution operation is a special linear operation that uses a filter (also called a convolution kernel) to slide over the input data and calculate the weighted sum of the data in each local area, thereby extracting the local features of the input process parameters. Therefore, adding a fully connected layer 225 after the first convolution layer 221, the second convolution layer 222, the third convolution layer 223, and the fourth convolution layer 224 can integrate the feature information of all previous layers and perform global classification or regression, thereby giving the model stronger expressive power.

[0049] The convolution kernel size of each first filter in the first convolution layer 221, the convolution kernel size of each second filter in the second convolution layer 222, the convolution kernel size of each third filter in the third convolution layer 223, and the convolution kernel size of each fourth filter in the fourth convolution layer 224 can be set according to actual needs.

[0050] In an exemplary embodiment, the convolution kernel size of each first filter in the first convolution layer 221, the convolution kernel size of each second filter in the second convolution layer 222, the convolution kernel size of each third filter in the third convolution layer 223, and the convolution kernel size of each fourth filter in the fourth convolution layer 224 are respectively 1*3.

[0051] In an exemplary embodiment, the intermediate layer further includes an activation function layer (not shown) located between adjacent two convolutional layers among the first convolutional layer, the second convolutional layer, the third convolutional layer, and the fourth convolutional layer. Accordingly, the step of extracting the features of the input process parameters via the intermediate layer further includes: using the activation function layer to perform nonlinear transformation operations on the first feature data of the M channels, the second feature data of the M channels, and the third feature data of the M channels.

[0052] Convolution is a linear operation, but neural networks need to fit nonlinear functions. Therefore, an activation function layer is added after the convolution layer to introduce nonlinearity, allowing the neural network to fit complex data distributions. Common activation functions include the sigmoid activation function, the rectified linear unit (ReLU), and the hyperbolic tangent activation function (Tanh).

[0053] In an exemplary embodiment, the activation function layer is a ReLU function layer, which has the advantages of simple calculation, fast convergence speed, and can effectively alleviate the vanishing gradient problem.

[0054] In an exemplary embodiment, the intermediate layer further includes a pooling layer and a batch normalization (BN) layer located after the activation function layer. Accordingly, the step of extracting the features of the input process parameters via the intermediate layer further includes: using the pooling layer to pool the multiple transformed features output by the activation function layer to obtain pooled input features; and using the batch normalization layer to perform batch normalization on the pooled input features.

[0055] The pooling layer, located after the convolutional layer, is used to perform dimensionality reduction and feature selection on the feature data output by the convolutional layer. Specifically, the pooling operation of the pooling layer aggregates local regions of the feature data using certain rules (such as maximum value, average value, etc.), thereby reducing the size of the feature data and reducing the computational effort. Pooling also improves the robustness and generalization ability of neural network models.

[0056] The batch normalization layer performs some form of "normalization" on the activation values ​​of each layer, which can maintain a relatively stable distribution of each layer's input during training. That is, the mean and variance do not fluctuate dramatically within a short training iteration. This normalization operation helps to alleviate gradient vanishing and gradient exploding problems. It also eliminates the need for the convolutional neural network system to constantly adapt to rapidly changing activation distributions, thereby improving learning efficiency and shortening training convergence time.

[0057] Please continue to refer to Figure 1 and Figure 2 , executing step S130, outputting the electrical parameter prediction result for the input process parameters via the output layer.

[0058] In an exemplary embodiment, the input process parameters are process parameters of a CMOS device. Accordingly, the output layer 230 outputs electrical parameter prediction results for the input process parameters, including one or more of threshold voltage, on-state current, off-state leakage current, transconductance, on-resistance, parasitic capacitance, gate oxide breakdown voltage, and source-drain breakdown voltage.

[0059] Generally speaking, data processing using convolutional neural networks can be divided into two phases: training and data processing. The training phase involves first training the neural network with training data to adjust the weights (also known as parameters) of the convolutional neural network. The data processing phase involves subsequently using the trained convolutional neural network to extract features from the input data and perform other tasks such as identifying and classifying objects within the data.

[0060] In an exemplary embodiment, the training process of a convolutional neural network system for electrical parameter prediction includes: a data collection phase, dividing the sample data set obtained from the production line into a training set and a validation set according to a preset ratio; a data training phase, using the training set to train the fully connected convolutional neural network system separately until the fully connected convolutional neural network system reaches convergence on the preset validation set.

[0061] In an exemplary embodiment, a preset number of training data sets are used each time to perform an iterative training on a convolutional neural network system with initial weights, thereby completing an adjustment of the weights of the convolutional neural network system.

[0062] Specifically, the process of each iterative training includes: using a preset number of training data to train the convolutional neural network system to be trained for electrical parameter prediction, and obtaining a preset number of training results; according to the difference between the training results and the corresponding correct results, respectively using a preset loss function to calculate the loss value; performing network back propagation derivation based on the calculated loss value to obtain the gradient value; according to the gradient value obtained by back propagation derivation, adjusting the weight of the convolutional neural network system for electrical parameter prediction.

[0063] Accordingly, multiple iterations of training are performed, and the weights of the convolutional neural network system for electrical parameter prediction are continuously iteratively updated until the loss value of the convolutional neural network system for electrical parameter prediction on the preset validation set converges. The convergence of the loss value of the convolutional neural network system for electrical parameter prediction on the preset second validation set means that the loss value of the convolutional neural network system for electrical parameter prediction on the preset second validation set reaches a minimum value.

[0064] For more detailed information about the multiple iterative training process of the convolutional neural network system for electrical parameter prediction, please refer to the iterative training process of the convolutional neural network system in the prior art, which will not be repeated here.

[0065] Accordingly, an embodiment of the present invention further provides a process parameter prediction device.

[0066] Figure 4 The schematic diagram of the structure of an embodiment of the process parameter prediction device provided by the technical solution of the present invention is shown. Figure 4, an electrical parameter prediction device 40 using a convolutional neural network system, the convolutional neural network system includes an input layer, an intermediate layer and an output layer, the electrical parameter prediction device 40 includes: a process parameter input unit 401, used to obtain the input process parameters through the input layer; a parameter feature extraction unit 402, used to extract the features of the input process parameters through the intermediate layer; a processing result output unit 403, used to output the electrical parameter prediction results for the input process parameters through the output layer.

[0067] The electrical parameter prediction device using a convolutional neural network system in the embodiment of the present invention can be used to execute the aforementioned electrical parameter prediction method using a convolutional neural network system, or other functional modules can be used to execute the aforementioned electrical parameter prediction method using a convolutional neural network system. For the electrical parameter prediction method using a convolutional neural network system provided in the embodiment of the present invention, please refer to the detailed description in the previous section and will not be repeated here.

[0068] Accordingly, an embodiment of the present invention also provides a computer device, which can implement the electrical parameter prediction method using a convolutional neural network system provided by an embodiment of the present invention by loading the above-mentioned electrical parameter prediction method using a convolutional neural network system in the form of a program.

[0069] refer to Figure 5 , which shows an optional hardware structure diagram of a computer device provided in an embodiment of the present invention. The computer device in the embodiment of the present invention includes: at least one processor 01, at least one communication interface 02, at least one memory 03 and at least one communication bus 04.

[0070] In this embodiment, the number of each of the processor 01 , the communication interface 02 , the memory 03 and the communication bus 04 is at least one, and the processor 01 , the communication interface 02 and the memory 03 communicate with each other via the communication bus 04 .

[0071] The communication interface 02 may be an interface of a communication module for network communication, such as an interface of a GSM module.

[0072] The processor 01 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the electrical parameter prediction method using a convolutional neural network system of this embodiment.

[0073] Memory 03 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage. Memory 03 stores one or more computer instructions, which are executed by processor 01 to implement the electrical parameter prediction method using a convolutional neural network system provided in the aforementioned embodiment.

[0074] It should be noted that the above-mentioned computer device may also include other devices (not shown) that may not be necessary for understanding the contents disclosed in the embodiments of the present invention; since these other devices may not be necessary for understanding the contents disclosed in the embodiments of the present invention, the embodiments of the present invention will not introduce them one by one.

[0075] Accordingly, an embodiment of the present invention also provides a computer program product, including a computer program / instruction, which, when executed by a processor, is used to implement the electrical parameter prediction method using a convolutional neural network system described in an embodiment of the present invention.

[0076] An embodiment of the present invention also provides a storage medium, which stores one or more computer instructions, and the one or more computer instructions are used to implement the electrical parameter prediction method using a convolutional neural network system provided in the above embodiment.

[0077] The embodiments of the present invention described above are combinations of elements and features of the present invention. Unless otherwise mentioned, elements or features may be considered as optional. Each element or feature may be put into practice without being combined with other elements or features. In addition, embodiments of the present invention may be constructed by combining some elements and / or features. The order of operations described in the embodiments of the present invention may be rearranged. Some configurations of any one embodiment may be included in another embodiment and may be replaced by the corresponding configuration of another embodiment. It is obvious to those skilled in the art that claims that do not have a clear reference relationship to each other in the appended claims may be combined into embodiments of the present invention, or may be included as new claims in amendments after submitting this application.

[0078] The embodiments of the present invention may be implemented by various means such as hardware, firmware, software, or a combination thereof. In a hardware configuration, the method according to the exemplary embodiment of the present invention may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, etc.

[0079] In a firmware or software configuration, the embodiments of the present invention may be implemented in the form of modules, procedures, functions, and the like. Software codes may be stored in a memory unit and executed by a processor. The memory unit may be located inside or outside the processor and may send and receive data to and from the processor via various known means.

[0080] The above description of the disclosed embodiments will enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

[0081] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be based on the scope defined by the claims.

Claims

1. A method for predicting electrical parameters using a convolutional neural network system, characterized in that: The convolutional neural network system includes an input layer, an intermediate layer, and an output layer. The electrical parameter prediction method includes: obtaining the input process parameters via the input layer; extracting features of the input process parameters via the intermediate layer; The electrical parameter prediction result for the input process parameter is outputted via the output layer.

2. The electrical parameter prediction method using a convolutional neural network system according to claim 1, wherein: The intermediate layer includes a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer and a fully connected layer, the first convolutional layer includes M first filters, the second convolutional layer includes M second filters, the third convolutional layer includes M third filters, and the fourth convolutional layer includes M fourth filters; The extracting of the features of the input process parameters through the intermediate layer includes: using the M first filters in the first convolutional layer to respectively receive the input electrical parameters, and performing a first convolution operation on the input electrical parameters to obtain first feature data of M channels; using the M second filters in the second convolutional layer to respectively receive the first feature data of the M channels, and performing a second convolution operation on the first feature data of the M channels to obtain second feature data of the M channels; using the M third filters in the third convolutional layer to respectively receive the second feature data of the M channels, and performing a third convolution operation on the second feature data of the M channels to obtain third feature data of the M channels; using the M fourth filters in the fourth convolutional layer to respectively receive the third feature data of the M channels, and performing a fourth convolution operation on the third feature data of the M channels to obtain fourth feature data of the M channels; using the fully connected layer to receive the fourth feature data of the M channels, and performing a fully connected operation on the fourth feature data of the M channels to obtain electrical parameter prediction results for the input process parameters.

3. The electrical parameter prediction method using a convolutional neural network system according to claim 2, wherein: The convolution kernel size of the first filter, the convolution kernel size of the second filter, the convolution kernel size of the third filter, and the convolution kernel size of the fourth filter are respectively 1*3.

4. The electrical parameter prediction method using a convolutional neural network system according to claim 2, wherein: The intermediate layer further includes an activation function layer located between two adjacent convolutional layers among the first convolutional layer, the second convolutional layer, the third convolutional layer, and the fourth convolutional layer; The extracting the features of the input process parameters via the intermediate layer also includes: using the activation function layer to perform nonlinear conversion operations on the first feature data of M channels, the second feature data of M channels, and the third feature data of M channels respectively.

5. The electrical parameter prediction method using a convolutional neural network system according to claim 4, wherein: The activation function layer includes a ReLU function layer.

6. The electrical parameter prediction method using a convolutional neural network system according to claim 4, wherein: The intermediate layer further includes at least one of a pooling layer and a batch normalization layer located after the activation function layer; The extracting the features of the input process parameters via the intermediate layer further includes: performing pooling processing on the multiple transformation features output by the activation function layer using the pooling layer to obtain pooled input features; The batch normalization layer is used to perform batch normalization processing on the pooled input features.

7. An electrical parameter prediction device using a convolutional neural network system, characterized in that: The convolutional neural network system includes an input layer, an intermediate layer, and an output layer. The electrical parameter prediction device includes: a process parameter input unit, configured to obtain the input process parameters via the input layer; a parameter feature extraction unit, configured to extract features of the input process parameters via the intermediate layer; The processing result output unit is used to output the electrical parameter prediction result for the input process parameter via the output layer.

8. A computer device, characterized in that: The method comprises at least one memory and at least one processor, wherein the memory stores one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of claims 1 to 6.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, they are used to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of claims 1 to 6.

10. A storage medium, characterized in that: The storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the electrical parameter prediction method using a convolutional neural network system as described in any one of claims 1 to 6.

Images