Process parameter prediction method and related equipment

Abstract

The invention discloses a process parameter prediction method and related equipment, and the method comprises the steps: obtaining an input electrical parameter; and performing process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under the various selectable process categories. According to the technical scheme provided by the embodiment of the invention, the corresponding relationship between the process parameters and the electrical parameters can be found under the condition of considering the process deviation, so that the problem of electrical parameter deviation caused by the process deviation in the actual manufacturing process of the semiconductor device can be avoided, and the product yield can be improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of semiconductor manufacturing, and in particular to a process parameter prediction method and related equipment. Background Art

[0002] In recent years, with the vigorous development of deep neural networks, academia and industry have witnessed major breakthroughs in deep learning in many fields.

[0003] The rapid development of fields such as information science, energy, and national defense has led to diverse demands on semiconductor devices. Based on currently mature processes, neural networks offer an option for rapidly predicting device structures or properties, ensuring accuracy while reducing computational costs and shortening development cycles.

[0004] Currently, the solution of using a neural network system to predict the electrical parameters of semiconductor devices using process parameters as input cannot consider the problem of electrical parameter offset caused by process deviation, thereby reducing product yield. Summary of the Invention

[0005] The problem solved by the embodiments of the present invention is to provide a process parameter prediction method and related equipment, which can find the correspondence between process parameters and electrical parameters while taking process deviation into account, thereby avoiding the electrical parameter offset problem caused by process deviation in the actual manufacturing process of semiconductor devices.

[0006] To solve the above problems, an embodiment of the present invention provides a process parameter prediction method, comprising:

[0007] Get input electrical parameters;

[0008] The input electrical parameters are processed by a preset neural network system to predict process parameters and obtain process parameter prediction results under multiple optional process categories.

[0009] Optionally, the neural network system includes a parameter sharing layer, a process parameter feature extraction layer, a process classification feature extraction layer, a first fully connected layer, a normalized exponential function layer, and a gate;

[0010] The input electrical parameters are processed for process parameter prediction via a preset neural network system to obtain process parameters under a variety of optional process categories, including: performing parameter sharing feature extraction on the input electrical parameters via the parameter sharing layer to obtain corresponding parameter sharing features; extracting corresponding process classification features from the parameter sharing features via the process classification feature extraction layer; performing a first fully connected operation on the process classification features via the first fully connected layer to obtain a process category prediction probability; performing a normalization operation on the process category prediction probability via the normalized exponential function layer to obtain a process category prediction result, which includes a variety of optional process categories; passing the process category prediction result to the process parameter feature extraction layer via a gate; performing process parameter prediction processing on the parameter sharing features based on the process category prediction result via the process parameter feature extraction layer to obtain process parameter prediction results under the various optional process categories.

[0011] Optionally, the parameter sharing layer, the process parameter feature extraction layer and the process classification feature extraction layer include at least one network block;

[0012] The at least one network block includes a cascaded second fully connected layer and an activation function layer; wherein the second fully connected layer receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer performs a nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

[0013] Optionally, the gate includes a ReLU activation function layer.

[0014] Optionally, dimensions of the process parameters in the process parameter prediction results under the multiple optional process categories are different.

[0015] Accordingly, an embodiment of the present invention further provides a process parameter prediction device, comprising:

[0016] An acquisition unit, used for acquiring input electrical parameters;

[0017] The prediction unit is used to perform process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0018] Optionally, the neural network system includes a parameter sharing layer, a process parameter feature extraction layer, a process classification feature extraction layer, a first fully connected layer, a normalized exponential function layer, and a gate;

[0019] The prediction unit is used to perform parameter sharing feature extraction on the input electrical parameters via the parameter sharing layer to obtain corresponding parameter sharing features; extract corresponding process classification features from the parameter sharing features via the process classification feature extraction layer; perform a first fully connected operation on the process classification features via the first fully connected layer to obtain a process category prediction probability; perform a normalization operation on the process category prediction probability via the normalized exponential function layer to obtain a process category prediction result, wherein the process category prediction result includes multiple optional process categories; pass the process category prediction result to the process parameter feature extraction layer via a gate; perform process parameter prediction processing on the parameter sharing features based on the process category prediction result via the process parameter feature extraction layer to obtain process parameter prediction results under the multiple optional process categories.

[0020] Optionally, the parameter sharing layer, the process parameter feature extraction layer and the process classification feature extraction layer include at least one network block;

[0021] The at least one network block includes a cascaded second fully connected layer and an activation function layer; wherein the second fully connected layer receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer performs a nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

[0022] Optionally, the gate includes a ReLU activation function layer.

[0023] Optionally, dimensions of the process parameters in the process parameter prediction results under the multiple optional process categories are different.

[0024] Correspondingly, 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 stores one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the process parameter prediction method as described in any one of the above items.

[0025] Accordingly, an embodiment of the present invention further provides a computer program product, including a computer program / instruction, which is used to implement any of the process parameter prediction methods described above when executed by a processor.

[0026] Correspondingly, an embodiment of the present invention further provides a storage medium, wherein the storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the process parameter prediction method as described in any one of the lines.

[0027] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0028] The process parameter prediction method provided by the embodiment of the present invention includes: obtaining input electrical parameters; performing process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0029] The process parameter prediction method provided by an embodiment of the present invention performs process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under a variety of different process categories. In other words, the neural network system can obtain situations where a variety of different electrical parameters correspond to the same process parameters under the same process category due to process deviations. The correspondence between process parameters and electrical parameters can be found while considering process deviations, thereby avoiding the problem of electrical parameter offset caused by process deviations in the actual manufacturing process of semiconductor devices, which is conducive to improving product yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of an embodiment of a process parameter prediction method provided by the technical solution of the present invention;

[0031] Figure 2 It is a structural diagram of an embodiment of a connected neural network system in the process parameter prediction method provided by the technical solution of the present invention;

[0032] Figure 3 It is a structural diagram of an embodiment of an intermediate layer connected to a neural network system in a process parameter prediction method provided by the technical solution of the present invention;

[0033] Figure 4 This is a structural diagram of an embodiment of a network block of a fully connected neural network system provided by the technical solution of the present invention;

[0034] Figure 5 It is a structural diagram of an embodiment of a process parameter prediction device provided by the technical solution of the present invention;

[0035] Figure 6 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

[0036] As can be seen from the background art, the solution of using a neural network system to predict the electrical parameters of semiconductor devices using process parameters as input cannot consider the problem of electrical parameter offset caused by process deviation, resulting in a decrease in product yield.

[0037] In order to solve the above technical problems, an embodiment of the present invention provides a process parameter prediction method, including: obtaining input electrical parameters; performing process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0038] The process parameter prediction method provided by an embodiment of the present invention performs process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under a variety of different process categories. In other words, the neural network system can obtain situations where a variety of different electrical parameters correspond to the same process parameters under the same process category due to process deviations. The correspondence between process parameters and electrical parameters can be found while considering process deviations, thereby avoiding the problem of electrical parameter offset caused by process deviations in the actual manufacturing process of semiconductor devices, which is conducive to improving product yield.

[0039] 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.

[0040] Figure 1 The following is a flow chart showing an embodiment of a process parameter prediction method provided by the technical solution of the present invention. Figure 1 A process parameter prediction method may include the following steps:

[0041] Step S110: obtaining input electrical parameters;

[0042] Step S120 : performing process parameter prediction processing on the input electrical parameters via a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0043] Please continue to refer to Figure 1 , execute step S110 to obtain input electrical parameters.

[0044] The input electrical parameters are obtained to provide a basis for subsequent process parameter prediction processing of the input electrical parameters through a preset neural network system, and to obtain process parameter prediction results under multiple optional process categories.

[0045] The input electrical parameters refer to expected values ​​or target values ​​of the electrical parameters of the semiconductor device, and can be set by the semiconductor device designer according to actual needs, and are not limited here.

[0046] In an exemplary embodiment, the input electrical parameters are electrical parameters of a complementary metal oxide semiconductor (CMOS) device. Specifically, the input electrical parameters include 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.

[0047] Please continue to refer to Figure 1 , executing step S120, performing process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0048] The input electrical parameters are processed for process parameter prediction through a preset neural network system to obtain process parameter prediction results under multiple optional process categories. That is, the neural network system can obtain situations where multiple different electrical parameters correspond to the same process parameters under the same process category due to process deviations, and can find the correspondence between process parameters and electrical parameters while considering process deviations, thereby avoiding the problem of electrical parameter offset caused by process deviations in the actual manufacturing process of semiconductor devices, which is conducive to improving product yield.

[0049] In an exemplary embodiment, the input electrical parameters are electrical parameters of a CMOS device, and accordingly, the process parameters are process parameters of a preparation process of a CMOS device, which may specifically 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 injection process, process parameters of a spacer formation process, process parameters of a source / drain injection process, and process parameters of an annealing process.

[0050] In one exemplary embodiment, the neural network system is a fully connected neural network system. A fully connected neural network is a classic feedforward neural network structure, typically composed of multiple stacked layers of neurons, with each layer of neurons connected to all neurons in the previous layer. This allows the fully connected neural network to perform complex feature extraction and nonlinear transformations, giving the model stronger expressive power.

[0051] Reference Figure 2The fully connected neural network system described includes an input layer 210, an intermediate layer 220, and an output layer 230. The input layer 210 is used to obtain the input electrical parameters, the intermediate layer 220 is used to extract process parameter features under multiple optional process categories based on the input electrical parameters, and the output layer 230 is used to output the process parameter features under the multiple optional process categories.

[0052] In an exemplary embodiment, referring to Figure 3 The described intermediate layer 220 may include: a parameter sharing layer 221 , a process classification feature extraction layer 222 , a first fully connected layer 223 , a normalized exponential function layer 224 , a gate 225 and a process parameter feature extraction layer 226 .

[0053] Accordingly, the process parameter prediction processing is performed on the input electrical parameters through the preset neural network system to obtain process parameters including multiple optional process categories, including: performing parameter sharing feature extraction on the input electrical parameters through the parameter sharing layer 221 to obtain corresponding parameter sharing features; extracting corresponding process classification features from the parameter sharing features through the process classification feature extraction layer 222; performing a first fully connected operation on the process classification features through the first fully connected layer 223 to obtain a process category prediction probability; performing a normalization operation on the process category prediction probability through the normalized exponential function layer 224 to obtain a process category prediction result, and the process category prediction result includes multiple optional process categories; passing the process category prediction result to the process parameter feature extraction layer 226 through the gate 225; performing process parameter prediction processing on the parameter sharing features based on the process category prediction result through the process parameter feature extraction layer 226 to obtain process parameter prediction results under the multiple optional process categories.

[0054] In one exemplary embodiment, the neural network system is a fully connected neural network system. Accordingly, the parameter sharing layer 221, the process parameter feature extraction layer 226, and the process classification feature extraction layer 222 are each fully connected neural network layers. Specifically, each of the parameter sharing layer 221, the process parameter feature extraction layer 226, and the process classification feature extraction layer 222 includes at least one network block 40.

[0055] In an exemplary embodiment, referring to Figure 4 The network block 40 described herein includes a cascaded second fully connected layer 410 and an activation function layer 420. The second fully connected layer 410 receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer 420 performs a nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

[0056] The second fully connected layer 410 performs a weighted sum operation on the input from the neurons in the previous layer and adds a bias term, and then the activation function layer 420 performs a nonlinear conversion operation on the fully connected features from the second fully connected layer 410, so that the parameter sharing layer, the process parameter feature extraction layer and the process classification feature extraction layer can capture the nonlinear relationship in the data and improve the expressive ability of the neural network.

[0057] Common activation functions include Sigmoid activation function, ReLU (Rectified Linear Unit), Tanh activation function, etc. In an exemplary embodiment, the activation function layer 320 is a ReLU activation function.

[0058] The gate 225 is used to control the extraction of process parameter features by the process parameter feature extraction layer 226. Specifically, the gate 22 is used to transmit the process category prediction result output by the normalized exponential function layer 224 to the process parameter feature extraction layer 226, so as to control the process parameter feature extraction layer 226 to activate neurons related to process parameters under the corresponding multiple optional process categories, while setting the weights of neurons related to process parameters under the multiple optional process categories that do not belong to the output of the normalized exponential function layer 224 to zero.

[0059] In other words, the process parameter feature extraction layer 226 can be regarded as a collection of process parameter prediction models applicable to different process categories, and the gate 225 is used to control the process parameter feature extraction layer 226 to respectively use the process parameter prediction models under multiple optional process categories in the process category prediction results output by the normalized exponential function layer 224 to perform process parameter feature extraction on the parameter sharing features extracted by the parameter sharing layer 221, thereby obtaining the process parameter prediction results under the multiple optional process categories.

[0060] In an exemplary embodiment, the gate 225 is a ReLU activation function.

[0061] In an exemplary embodiment, the fully connected neural network system can be used to predict process parameters under a variety of different process categories. Accordingly, the process parameter dimension is selected as the union of the process parameters under all process categories. Moreover, the dimension of the process parameters under each process category can be different, that is, the fully connected neural network model in this application can be used in situations where the process parameter dimension is not fixed, which is conducive to expanding the scope of application of the process parameter prediction of the connected neural network system and meeting the diverse process parameter prediction needs of users.

[0062] Generally speaking, data processing using neural network systems can be divided into two phases: training and data processing. The training phase involves first training the neural network using training data to adjust the neural network's weights (also known as parameters). The data processing phase involves subsequently using the trained neural network to extract features from the input data and perform other tasks such as identifying and classifying objects within the data.

[0063] In an exemplary embodiment, the training process of a neural network system for process 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 neural network system separately until the fully connected neural network system reaches convergence on the preset validation set.

[0064] In an exemplary embodiment, a preset number of training data sets are used each time to perform an iterative training on the neural network system with initial weights, thereby completing an adjustment of the weights of the neural network system.

[0065] Specifically, the process of each iterative training includes: using a preset number of training data to train the neural network system to be trained for process 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; based on the calculated loss value, performing network back propagation derivation to obtain the gradient value; based on the gradient value obtained by back propagation derivation, adjusting the weight of the neural network system for process parameter prediction.

[0066] Accordingly, multiple iterations of training are performed, and the weights of the neural network system for process parameter prediction are continuously updated iteratively until the loss value of the neural network system for process parameter prediction on the validation set converges. The convergence of the loss value of the neural network system for process parameter prediction on the validation set means that the loss value of the neural network system for process parameter prediction on the validation set reaches a minimum.

[0067] For more detailed information about the multiple iterative training process of the neural network system for process parameter prediction, please refer to the prior art on the iterative training process of the neural network system, which will not be repeated here.

[0068] Accordingly, an embodiment of the present invention further provides a process parameter prediction device.

[0069] Figure 5 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 5A process parameter prediction device 50 may include: an acquisition unit 501 for acquiring input electrical parameters; a prediction unit 502 for performing process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

[0070] In an exemplary embodiment, the neural network system includes a parameter sharing layer, a process parameter feature extraction layer, a process classification feature extraction layer, a first fully connected layer, a normalized exponential function layer, and a gate. Accordingly, the prediction unit 502 is configured to perform parameter sharing feature extraction on the input electrical parameters via the parameter sharing layer to obtain corresponding parameter sharing features; extract corresponding process classification features from the parameter sharing features via the process classification feature extraction layer; perform a first fully connected operation on the process classification features via the first fully connected layer to obtain a process category prediction probability; perform a normalization operation on the process category prediction probability via the normalized exponential function layer to obtain a process category prediction result, wherein the process category prediction result includes multiple optional process categories; transmit the process category prediction result to the process parameter feature extraction layer via the gate; perform process parameter prediction processing on the parameter sharing features based on the process category prediction result via the process parameter feature extraction layer to obtain process parameter prediction results under the multiple optional process categories.

[0071] In an exemplary embodiment, the parameter sharing layer, the process parameter feature extraction layer and the process classification feature extraction layer include at least one network block, and the at least one network block includes a cascaded second fully connected layer and an activation function layer; wherein the second fully connected layer receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer performs nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

[0072] In an exemplary embodiment, the activation function is ReLU.

[0073] In an exemplary embodiment, dimensions of the process parameters in the process parameter prediction results under the multiple selectable process categories are different.

[0074] The process parameter prediction device in the embodiment of the present invention can be used to execute the aforementioned process parameter prediction method, or other functional modules can be used to execute the aforementioned process parameter prediction method. For the process parameter prediction method provided by the embodiment of the present invention, please refer to the detailed description of the aforementioned part, which will not be repeated here.

[0075] Accordingly, an embodiment of the present invention further provides a computer device, which can implement the process parameter prediction method provided by the embodiment of the present invention in the form of a loaded program.

[0076] refer to Figure 6 , 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.

[0077] 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 .

[0078] The communication interface 02 may be an interface of a communication module for network communication, such as an interface of a GSM module.

[0079] The processor 01 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the process parameter prediction method of this embodiment.

[0080] The memory 03 may include a high-speed RAM memory, or may also include a non-volatile memory, such as at least one disk storage. The memory 03 stores one or more computer instructions, which are executed by the processor 01 to implement the process parameter prediction method provided in the aforementioned embodiment.

[0081] It should be noted that the above-mentioned electronic 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.

[0082] Accordingly, an embodiment of the present invention further provides a computer program product, including a computer program / instruction, which is used to implement the process parameter prediction method described in the embodiment of the present invention when executed by a processor.

[0083] An embodiment of the present invention further provides a storage medium, wherein the storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the process parameter prediction method provided in the above embodiment.

[0084] 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.

[0085] 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.

[0086] 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.

[0087] 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.

[0088] 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 process parameter prediction method, characterized in that: include: Get input electrical parameters; The input electrical parameters are processed by a preset neural network system to predict process parameters and obtain process parameter prediction results under multiple optional process categories.

2. The process parameter prediction method according to claim 1, wherein: The neural network system includes a parameter sharing layer, a process parameter feature extraction layer, a process classification feature extraction layer, a first fully connected layer, a normalized exponential function layer and a gate; The input electrical parameters are processed by a preset neural network system to predict process parameters, and process parameters under multiple optional process categories are obtained, including: performing parameter sharing feature extraction on the input electrical parameters via the parameter sharing layer to obtain corresponding parameter sharing features; extracting corresponding process classification features from the parameter sharing features via the process classification feature extraction layer; performing a first fully connected operation on the process classification features via the first fully connected layer to obtain a process category prediction probability; performing a normalization operation on the process category prediction probability via the normalized exponential function layer to obtain a process category prediction result, and the process category prediction result includes multiple optional process categories; passing the process category prediction result to the process parameter feature extraction layer via the gate; performing process parameter prediction processing on the parameter sharing features based on the process category prediction result via the process parameter feature extraction layer to obtain process parameter prediction results under the multiple optional process categories.

3. The process parameter prediction method according to claim 2, wherein: The parameter sharing layer, the process parameter feature extraction layer, and the process classification feature extraction layer each include at least one network block; The at least one network block includes a cascaded second fully connected layer and an activation function layer; wherein the second fully connected layer receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer performs a nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

4. The process parameter prediction method according to claim 2, wherein: The gate includes a ReLU activation function layer.

5. The process parameter prediction method according to claim 2, wherein: The dimensions of the process parameters in the process parameter prediction results under the multiple optional process categories are different.

6. A process parameter prediction device, characterized in that: include: An acquisition unit, used for acquiring input electrical parameters; The prediction unit is used to perform process parameter prediction processing on the input electrical parameters through a preset neural network system to obtain process parameter prediction results under multiple optional process categories.

7. The process parameter prediction device according to claim 6, characterized in that: The neural network system includes a parameter sharing layer, a process parameter feature extraction layer, a process classification feature extraction layer, a first fully connected layer, a normalized exponential function layer and a gate; The prediction unit is configured to perform parameter sharing feature extraction on the input electrical parameters via the parameter sharing layer to obtain corresponding parameter sharing features; and extract corresponding process classification features from the parameter sharing features via the process classification feature extraction layer; Performing a first fully connected operation on the process classification feature via the first fully connected layer to obtain a process category prediction probability; performing a normalization operation on the process category prediction probability via a normalized exponential function layer to obtain a process category prediction result, wherein the process category prediction result includes a plurality of optional process categories; passing the process category prediction result to the process parameter feature extraction layer via a gate; The process parameter feature extraction layer performs process parameter prediction processing on the parameter sharing feature based on the process category prediction result to obtain process parameter prediction results under the multiple optional process categories.

8. The process parameter prediction device according to claim 7, characterized in that: The parameter sharing layer, the process parameter feature extraction layer, and the process classification feature extraction layer each include at least one network block; The at least one network block includes a cascaded second fully connected layer and an activation function layer; wherein the second fully connected layer receives multiple input features and performs a second fully connected operation on the received multiple input features to obtain corresponding multiple fully connected features; the activation function layer performs a nonlinear transformation on the multiple fully connected features to obtain corresponding transformed features.

9. The process parameter prediction device according to claim 7, wherein: The gate includes a ReLU activation function layer.

10. The process parameter prediction device according to claim 7, wherein: The dimensions of the process parameters in the process parameter prediction results under the multiple optional process categories are different.

11. 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 process parameter prediction method according to any one of claims 1 to 5.

12. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, it is used to implement the process parameter prediction method according to any one of claims 1 to 5.

13. 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 process parameter prediction method according to any one of claims 1 to 5.

Images