An application code development method and system based on artificial intelligence model

Abstract

An embodiment of the present invention provides an application code development method and system based on an artificial intelligence model, including: collecting network demand data; classifying the network demand data to obtain demand classification samples of different dimensions; inputting the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and the corresponding low-code development data; acquiring low-code development data according to the demand classification samples; dividing the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, and inputting the training set into a low-code small sample model to obtain a trained low-code small sample model; inputting new demand data into the trained low-code small sample model to obtain output specific low-code development data; realizing the expansion of application demand samples, so that user demand data matches application code development data, and improving application code development efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an application code development method based on an artificial intelligence model, an application code development system based on an artificial intelligence model, a computer device and a computer-readable storage medium. Background Art

[0002] Existing application development generally includes requirements analysis, system design, detailed design, code implementation, unit testing, integration testing, system testing, system deployment and maintenance. These steps constitute the entire process of application development. Each stage is crucial to ensure the quality of the final product and user satisfaction.

[0003] In addition, existing application code development can be developed through a low-code development platform. Compared with traditional application code development, low-code development has the following advantages: high versatility, low cost, and high connectivity. However, current low-code development has the problem of inaccurate development functions, which makes the developed functions not fully matched with the needs. Summary of the invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide an application code development method based on an artificial intelligence model, an application code development system based on an artificial intelligence model, a computer device and a computer-readable storage medium that overcome the above problems or at least partially solve the above problems.

[0005] In order to solve the above problems, an embodiment of the present invention discloses an application code development method based on an artificial intelligence model, the method comprising:

[0006] Collect network demand data;

[0007] Classifying the network demand data to obtain demand classification samples of different dimensions;

[0008] Obtain low-code development data based on the demand classification samples;

[0009] Input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and the corresponding low-code development data;

[0010] Divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into a low-code small sample model, and obtain a trained low-code small sample model;

[0011] Input the new requirement data into the trained low-code small sample model to obtain output specific low-code development data.

[0012] Preferably, the network demand data is classified to obtain demand classification samples of different dimensions, including:

[0013] The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

[0014] Preferably, extracting the demand features of the network demand data, classifying the network demand data according to the demand features, and obtaining demand classification samples includes:

[0015] Extracting industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data;

[0016] The network demand data is classified according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

[0017] Preferably, the inputting of the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain the generated virtual demand classification samples and the corresponding low-code development data includes:

[0018] Calculate the quantitative proportion of the industrial demand samples, agricultural demand samples, and service industry demand samples;

[0019] The number and proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model remain unchanged and are set as the conditional information of the generator network in the conditional generative adversarial network;

[0020] The generator network receives random noise and conditional information as input to generate virtual data that matches the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information to determine the authenticity of the virtual data, and obtains output virtual demand classification samples and corresponding low-code development data.

[0021] Preferably, the virtual demand classification samples and the corresponding low-code development data are divided into a training set and a test set, and the training set is input into a low-code small sample model to obtain a trained low-code small sample model, including:

[0022] Input the training set into the feature convolution module to obtain output first dimension data;

[0023] Processing the first dimensional data through a residual module to obtain second dimensional data;

[0024] The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data;

[0025] Input the third dimension data into the concat operation module to obtain three feature data;

[0026] The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, the weight parameters of the support vector machine model module are adjusted, and the trained low-code small sample model is obtained.

[0027] Preferably, the step of inputting the training set into a feature convolution module to obtain output first-dimensional data includes:

[0028] The image data and text data of the training set are identified;

[0029] Extracting text content of the image data and converting it into first feature sequence data;

[0030] Converting the text data into second feature sequence data;

[0031] The first feature sequence data and the second feature sequence data are combined into feature sequence data, and the feature sequence data is input into a feature convolution module to obtain output first dimensional data.

[0032] The embodiment of the present invention discloses an application code development system based on an artificial intelligence model, the system comprising:

[0033] A collection module, used to collect network demand data;

[0034] A demand classification module, used to classify the network demand data to obtain demand classification samples of different dimensions;

[0035] A low-code development data module is used to obtain low-code development data according to the required classification samples;

[0036] A generation module, used to input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain a generated virtual demand classification sample and the corresponding low-code development data;

[0037] A training module, used to divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into the low-code small sample model, and obtain a trained low-code small sample model;

[0038] The output module is used to input new demand data into the trained low-code small sample model to obtain output specific low-code development data.

[0039] Preferably, the network demand data is classified to obtain demand classification samples of different dimensions, including:

[0040] The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

[0041] An embodiment of the present invention discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned application code development method based on the artificial intelligence model when executing the computer program.

[0042] An embodiment of the present invention discloses a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned application code development method based on the artificial intelligence model are implemented.

[0043] The embodiments of the present invention include the following advantages:

[0044] In an embodiment of the present invention, the application code development method based on the artificial intelligence model includes: collecting network demand data; classifying the network demand data to obtain demand classification samples of different dimensions; inputting the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and corresponding low-code development data; acquiring low-code development data according to the demand classification samples; dividing the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, and inputting the training set into a low-code small sample model to obtain a trained low-code small sample model; inputting new demand data into the trained low-code small sample model to obtain output specific low-code development data; realizing the expansion of application demand samples, improving demand sample data, achieving the technical effect of automatic generation of demand data, improving the efficiency of model training, and creating an artificial intelligence model suitable for small samples, so that user demand data matches application code development data, and improving application code development efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 is a schematic diagram of an embodiment of an application code development method based on an artificial intelligence model according to an embodiment of the present invention;

[0047] Figure 2 is a schematic diagram of a conditional generative adversarial network according to an embodiment of the present invention;

[0048] Figure 3 It is a structural block diagram of an embodiment of an application code development system based on an artificial intelligence model according to an embodiment of the present invention;

[0049] Figure 4 The diagram is a diagram of the internal structure of a computer device according to an embodiment. DETAILED DESCRIPTION

[0050] In order to make the technical problems, technical solutions and beneficial effects solved by the embodiments of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] In one core of an embodiment of the present invention, the lack of user demand data easily leads to a lack of samples of low-code development data. In an embodiment of the present invention, a small amount of demand samples are obtained, and then virtual demand samples are generated through an artificial intelligence model to expand the application demand samples, thereby improving the demand sample data, achieving the technical effect of automatic generation of demand data, and improving the efficiency of model training; on the other hand, in order to improve the accuracy of application code development generation, the YOLOv3 model and the support vector machine model are combined, and the pooling layer and the fully connected layer are not set. The size transformation is achieved by changing the step size of the convolution kernel, and then the output features are classified and processed, and an artificial intelligence model suitable for small samples is created, so that the user demand data matches the application code development data, thereby improving the efficiency of application code development.

[0052] Reference Figure 1 , shows a schematic diagram of an embodiment of an application code development method based on an artificial intelligence model according to an embodiment of the present invention, which may specifically include the following steps:

[0053] Step S101, collecting network demand data;

[0054] The application code development method of the embodiment of the present invention can be applied to a terminal; the terminal may include various different terminals such as a personal computer, a laptop computer, a smart phone, a smart terminal, and a portable wearable device. The embodiment of the present invention does not limit the specific type of the terminal. The operating system of the terminal may include Android, Harmony OS, IOS, Windows Phone, Windows, etc., and the present invention does not impose too many restrictions on this.

[0055] In an embodiment of the present invention, first, network demand data can be collected. Specifically, the demand data on the network can be obtained in a variety of ways, that is, the network demand data can include a variety of user demand information on the network. Specifically, the demand data can refer to the description of the functions and feature modules that the application must have from the perspective of the end user so that the user can effectively complete the task or achieve the goal; specifically applied to an embodiment of the present invention, automated data collection can be performed through the HTTP protocol, API, web page structure and browser to obtain a variety of different demand data. In terms of specific data collection tools, data collection can be performed through a variety of tools such as BeautifulSoup, Scrapy, Selenium, etc. It should be noted that the network demand data is data associated with application development.

[0056] Step S102: classify the network demand data to obtain demand classification samples of different dimensions;

[0057] In an embodiment of the present invention, classifying the network demand data to obtain demand classification samples of different dimensions includes: extracting demand features of the network demand data, classifying the network demand data according to the demand features, and obtaining demand classification samples.

[0058] In a certain dimension, the demand feature may refer to different types of keywords under the same division dimension. For example, the demand data collected is divided into three different dimensions of industry, service industry, and agriculture; wherein the demand feature of the network demand data is extracted, and the network demand data is classified according to the demand feature to obtain a demand classification sample, including:

[0059] Extracting industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data;

[0060] The network demand data is classified according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

[0061] Specifically, the industrial dimension features can refer to industry-related keywords such as "machinery", "equipment", "chemical", "material", "energy", "coal", and "textile", while the agricultural dimension features can refer to keywords such as "planting", "forest", livestock Mu”, “Fisheries” and other agriculture-related keywords, while the service industry dimensional characteristics can refer to keywords related to the service industry such as “finance”, “leasing services”, “technical services”, “accommodation and catering”, “transportation”, “accommodation and catering”, “information transmission”, “education”, “health, social security”, “culture, sports and entertainment”, etc. Firstly, the dimensional characteristics of the network demand data are extracted, and according to the classification information of the pre-divided demand characteristics, the demand data corresponding to the demand characteristics are summarized into industrial demand samples, agricultural demand samples, and service industry demand samples.

[0062] It is worth noting that the above-mentioned division of dimensional features is an example of an embodiment of the present invention. The network demand data can also be divided according to other dimensional features, such as dividing the network demand data according to different user types (such as general users, corporate users, and public sector users). The embodiment of the present invention does not impose too many restrictions on the division method of dimensional features.

[0063] Step S103: acquiring low-code development data according to the demand classification samples;

[0064] In an embodiment of the present invention, the low-code development data may include application module drag and drop operation record data, application module layout and settings, logical association configuration and jump configuration of different application modules, command line record data, workflow data and built-in data flow, server connection deployment setting data, application lifecycle management data, etc. The embodiment of the present invention does not impose too many restrictions on this;

[0065] Regarding the method of obtaining the low-code development data, specifically, the above-mentioned demand classification samples can be input into the low-code development platform tool, and the control terminal records the platform tool records in each demand classification sample process, including the drag and drop operation record data of the application module, the application module layout and settings, the logical association configuration and jump configuration of different application modules, the command line record data, the workflow data and the built-in data flow, the server connection deployment setting data, the application lifecycle management data, etc.;

[0066] In a specific example, a requirement of the multiple types of samples may include: under the weather conditions of a certain quarter on the farm, calculate the proportion of each crop in a certain quarter in all crops on a certain farm and the estimated yield of each crop in the next quarter of the farm. After the control terminal inputs the requirement into the low-code development platform tool, each operation step of the low-code development platform tool obtains the drag operation record data of the above-mentioned application module, application module layout and settings, logical association configuration and jump configuration of different application modules, command line record data, workflow data and built-in data flow, server connection deployment setting data, application lifecycle management data, etc.

[0067] Step S104: input the demand classification sample and the corresponding low-code development data into a conditional generative adversarial network to obtain a generated virtual demand classification sample and the corresponding low-code development data;

[0068] Specifically applied to the embodiments of the present invention, after obtaining the demand classification samples and the corresponding low-code development data, the demand classification samples and the corresponding low-code development data can be input into the conditional generative adversarial network to generate virtual data, expand the demand sample data in quantity, and improve the accuracy of model training.

[0069] Specifically applied to the embodiment of the present invention, the demand classification sample and the corresponding low-code development data are input into the conditional generative adversarial network to obtain the generated virtual demand classification sample and the corresponding low-code development data, including:

[0070] Calculate the quantitative proportion of the industrial demand samples, agricultural demand samples, and service industry demand samples;

[0071] The number and proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model remain unchanged and are set as the conditional information of the generator network in the conditional generative adversarial network;

[0072] The generator network receives random noise and conditional information as input to generate virtual data that matches the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information to determine the authenticity of the virtual data, thereby obtaining output samples of multiple types of generated demand.

[0073] In the embodiment of the present invention, the number of the three demand samples in the total is first calculated, such as the proportion of industrial demand samples is 38%, the proportion of agricultural demand samples is 26%, and the proportion of industrial demand samples is 36%; the above-mentioned proportions are set as conditional information to improve the accuracy of the output virtual data. Figure 2 As shown, the conditional generative adversarial network may include a generator network and a discriminator network. The generator network receives random noise and condition information as input to generate virtual data matching the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and condition information to judge the authenticity of the virtual data. The generator network and the discriminator network are alternately iteratively trained to optimize the generator network and the discriminator network to obtain output virtual demand classification samples and corresponding low-code development data, thereby achieving the technical effect of automatic generation of demand data and code development data.

[0074] Step S105: divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into the low-code small sample model, and obtain a trained low-code small sample model;

[0075] Specifically applied to the embodiment of the present invention, the terminal can divide the virtual demand classification sample and the corresponding low-code development data into a training set and a test set. The training set is used to train the low-code small sample model, and the test set is used to test the accuracy of the low-code small sample model after training.

[0076] In a specific training process, the training set is input into the low-code small sample model to obtain a trained low-code small sample model, including:

[0077] Input the training set into the feature convolution module to obtain output first dimension data;

[0078] Processing the first dimensional data through a residual module to obtain second dimensional data;

[0079] The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data;

[0080] Input the third dimension data into the concat operation module to obtain three feature data;

[0081] The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, the weight parameters of the support vector machine model module are adjusted, and the trained low-code small sample model is obtained.

[0082] In an embodiment of the present invention, the low-code small sample model of the present application is realized by combining the YOLOv3 model and the support vector machine model. The advantage of the YOLOv3 model is that it has no pooling layer and fully connected layer. While maintaining the training speed, the detection accuracy of small samples is improved. Combined with the support vector machine, the classification ability of the model is improved, and the accuracy of model training is improved. In an embodiment of the present invention, the training set is input into the feature convolution module, the residual module, the upper pooling module, and the concat operation module to obtain three feature data, and then the feature data is transmitted to the support vector machine model module for iterative update to obtain the trained low-code small sample model. The above-mentioned model training process can be a training process set by technical personnel in this field according to actual conditions, and the embodiment of the present invention will not go into details one by one.

[0083] It should be noted that the YOLOv3 model may include a feature convolution module, a residual module, an upper pooling module, a concat operation module, etc., and of course, other modules. The embodiment of the present invention does not impose too many restrictions on the specific structure of the YOLOv3 model; the feature convolution module may refer to a combination of a convolution layer, a BN (Batch Normalization) layer and a Leaky ReLU (activation layer) layer, and the embodiment of the present invention does not impose too many restrictions on this; further applied to the embodiment of the present invention, the structure of the support vector machine model may include multiple types of support vector machines such as linear separable support vector machines, linear support vector machines, and nonlinear support vector machines, and the embodiment of the present invention does not impose too many restrictions on this.

[0084] In a preferred embodiment of the present invention, the loss function of the support vector machine model module can be as follows:

[0085] ;

[0086] in, represents the loss function, Indicates label data, the value is +1 or -1, represents the output of the decision function, represents the normal vector of the support vector machine model hyperplane, represents the input sample data, represents the bias term, Indicates the proportion coefficient of the first type of demand samples.

[0087] In a specific example, the inputting of the training set into the feature convolution module to obtain the output first dimension data includes: identifying the image data and text data of the training set; extracting the text content of the image data and converting it into first feature sequence data; converting the text data into second feature sequence data; merging the first feature sequence data and the second feature sequence data into feature sequence data, and inputting the feature sequence data into the feature convolution module to obtain the output first dimension data. Before inputting into the YOLOv3 model, the image data of the training set needs to be subjected to text extraction, and the extracted text and the original text data of the training set are converted into feature sequence data, and then the feature sequence data is input into the feature convolution module to obtain the output first dimension data.

[0088] Step S106: input the new demand data into the trained low-code small sample model to obtain output specific low-code development data.

[0089] Specifically applied to the embodiments of the present invention, after obtaining the trained low-code small sample model, new demand data can be input into the trained low-code small sample model to obtain output specific low-code development data, that is, the output low-code development platform tool characteristics, application module drag operation record data, application module layout and settings, logical association configuration and jump configuration of different application modules, command line record data, workflow data and built-in data flow, server connection deployment setting data, application lifecycle management data, etc. The embodiments of the present invention do not impose too many restrictions on this.

[0090] In an embodiment of the present invention, the application code development method based on the artificial intelligence model includes: collecting network demand data; classifying the network demand data to obtain demand classification samples of different dimensions; inputting the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and corresponding low-code development data; acquiring low-code development data according to the demand classification samples; dividing the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, and inputting the training set into a low-code small sample model to obtain a trained low-code small sample model; inputting new demand data into the trained low-code small sample model to obtain output specific low-code development data; realizing the expansion of application demand samples, improving demand sample data, achieving the technical effect of automatic generation of demand data, improving the efficiency of model training, and creating an artificial intelligence model suitable for small samples, so that user demand data matches application code development data, and improving application code development efficiency.

[0091] It should be noted that, for the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that this embodiment is not limited by the order of the actions described, because according to this embodiment, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by this embodiment.

[0092] Reference Figure 3 , shows a structural block diagram of an embodiment of an application code development system based on an artificial intelligence model of this embodiment, which may specifically include the following modules:

[0093] The collection module 301 is used to collect network demand data;

[0094] The demand classification module 302 is used to classify the network demand data to obtain demand classification samples of different dimensions;

[0095] A low-code development data module 303 is used to obtain low-code development data according to the demand classification samples;

[0096] A generation module 304 is used to input the demand classification sample and the corresponding low-code development data into a conditional generative adversarial network to obtain a generated virtual demand classification sample and the corresponding low-code development data;

[0097] A training module 305 is used to divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into the low-code small sample model, and obtain a trained low-code small sample model;

[0098] The output module 306 is used to input new demand data into the trained low-code small sample model to obtain output specific low-code development data.

[0099] Preferably, the network demand data is classified to obtain demand classification samples of different dimensions, including:

[0100] The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

[0101] Preferably, the demand classification module includes:

[0102] An extraction submodule, used to extract industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data;

[0103] The classification submodule is used to classify the network demand data according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

[0104] Preferably, the generating module comprises:

[0105] A calculation submodule, used to calculate the quantity ratio of the industrial demand samples, the agricultural demand samples, and the service industry demand samples;

[0106] Set the proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model to remain unchanged as the conditional information of the generator network in the conditional generative adversarial network;

[0107] The generation submodule is used to receive random noise and conditional information as input through the generator network, generate virtual data matching the conditions, receive industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information through the discriminator network, judge the authenticity of the virtual data, and obtain output virtual demand classification samples and corresponding low-code development data.

[0108] Preferably, the training module includes:

[0109] A first dimensional data submodule, used for inputting the training set into a feature convolution module to obtain output first dimensional data;

[0110] A second dimensional data submodule, used for processing the first dimensional data through a residual module to obtain second dimensional data;

[0111] A third dimensional data submodule, used for enlarging the size of the second dimensional data through an upper pooling module to obtain third dimensional data;

[0112] A feature data submodule, used for inputting the third dimension data into a concat operation module to obtain three feature data;

[0113] The optimization submodule is used to input the three feature data into the support vector machine model module, iteratively update, optimize the loss function of the support vector machine model module, adjust the weight parameters of the support vector machine model module, and obtain the trained low-code small sample model.

[0114] Preferably, the first dimension data submodule includes:

[0115] A recognition unit, used for recognizing the image data and text data of the training set;

[0116] An extraction unit, used for extracting text content of the image data and converting it into first feature sequence data;

[0117] A conversion unit, used for converting the text data into second feature sequence data;

[0118] The merging unit is used to merge the first feature sequence data and the second feature sequence data into feature sequence data, input the feature sequence data into a feature convolution module, and obtain output first dimensional data.

[0119] Each module in the above-mentioned application code development system based on artificial intelligence model can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0120] The above-mentioned artificial intelligence model-based application code development system can be used to execute the artificial intelligence model-based application code development method provided in any of the above-mentioned embodiments, and has corresponding functions and beneficial effects.

[0121] In one embodiment, a computer device is provided, whose internal structure diagram can be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for developing an application code system based on an artificial intelligence model is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0122] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0123] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the above embodiment are implemented:

[0124] Collect network demand data;

[0125] Classifying the network demand data to obtain demand classification samples of different dimensions;

[0126] Obtain low-code development data based on the demand classification samples;

[0127] Input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and the corresponding low-code development data;

[0128] Divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into a low-code small sample model, and obtain a trained low-code small sample model;

[0129] Input the new requirement data into the trained low-code small sample model to obtain output specific low-code development data.

[0130] Preferably, the network demand data is classified to obtain demand classification samples of different dimensions, including:

[0131] The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

[0132] Preferably, extracting the demand features of the network demand data, classifying the network demand data according to the demand features, and obtaining demand classification samples includes:

[0133] Extracting industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data;

[0134] The network demand data is classified according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

[0135] Preferably, the inputting of the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain the generated virtual demand classification samples and the corresponding low-code development data includes:

[0136] Calculate the quantitative proportion of the industrial demand samples, agricultural demand samples, and service industry demand samples;

[0137] The number and proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model remain unchanged and are set as the conditional information of the generator network in the conditional generative adversarial network;

[0138] The generator network receives random noise and conditional information as input to generate virtual data that matches the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information to determine the authenticity of the virtual data, and obtains output virtual demand classification samples and corresponding low-code development data.

[0139] Preferably, the virtual demand classification samples and the corresponding low-code development data are divided into a training set and a test set, and the training set is input into a low-code small sample model to obtain a trained low-code small sample model, including:

[0140] Input the training set into the feature convolution module to obtain output first dimension data;

[0141] Processing the first dimensional data through a residual module to obtain second dimensional data;

[0142] The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data;

[0143] Input the third dimension data into the concat operation module to obtain three feature data;

[0144] The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, the weight parameters of the support vector machine model module are adjusted, and the trained low-code small sample model is obtained.

[0145] Preferably, the step of inputting the training set into a feature convolution module to obtain output first-dimensional data includes:

[0146] The image data and text data of the training set are identified;

[0147] Extracting text content of the image data and converting it into first feature sequence data;

[0148] Converting the text data into second feature sequence data;

[0149] The first feature sequence data and the second feature sequence data are combined into feature sequence data, and the feature sequence data is input into a feature convolution module to obtain output first dimensional data.

[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above embodiment are implemented:

[0151] Collect network demand data;

[0152] Classifying the network demand data to obtain demand classification samples of different dimensions;

[0153] Obtain low-code development data based on the demand classification samples;

[0154] Input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and the corresponding low-code development data;

[0155] Divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into a low-code small sample model, and obtain a trained low-code small sample model;

[0156] Input the new requirement data into the trained low-code small sample model to obtain output specific low-code development data.

[0157] Preferably, the network demand data is classified to obtain demand classification samples of different dimensions, including:

[0158] The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

[0159] Preferably, extracting the demand features of the network demand data, classifying the network demand data according to the demand features, and obtaining demand classification samples includes:

[0160] Extracting industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data;

[0161] The network demand data is classified according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

[0162] Preferably, the inputting of the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain the generated virtual demand classification samples and the corresponding low-code development data includes:

[0163] Calculate the quantitative proportion of the industrial demand samples, agricultural demand samples, and service industry demand samples;

[0164] The number and proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model remain unchanged and are set as the conditional information of the generator network in the conditional generative adversarial network;

[0165] The generator network receives random noise and conditional information as input to generate virtual data that matches the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information to determine the authenticity of the virtual data, and obtains output virtual demand classification samples and corresponding low-code development data.

[0166] Preferably, the virtual demand classification samples and the corresponding low-code development data are divided into a training set and a test set, and the training set is input into a low-code small sample model to obtain a trained low-code small sample model, including:

[0167] Input the training set into the feature convolution module to obtain output first dimension data;

[0168] Processing the first dimensional data through a residual module to obtain second dimensional data;

[0169] The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data;

[0170] Input the third dimension data into the concat operation module to obtain three feature data;

[0171] The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, the weight parameters of the support vector machine model module are adjusted, and the trained low-code small sample model is obtained.

[0172] Preferably, the step of inputting the training set into a feature convolution module to obtain output first-dimensional data includes:

[0173] The image data and text data of the training set are identified;

[0174] Extracting text content of the image data and converting it into first feature sequence data;

[0175] Converting the text data into second feature sequence data;

[0176] The first feature sequence data and the second feature sequence data are combined into feature sequence data, and the feature sequence data is input into a feature convolution module to obtain output first dimensional data.

[0177] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0178] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0179] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks 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 terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal 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.

[0180] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including 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.

[0181] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement 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.

[0182] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0183] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0184] The above is a detailed introduction to an application code development method based on an artificial intelligence model, an application code development system based on an artificial intelligence model, a computer device and a computer-readable storage medium provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for developing application code based on an artificial intelligence model, characterized in that: The method comprises: Collect network demand data; Classifying the network demand data to obtain demand classification samples of different dimensions; Obtain low-code development data based on the demand classification samples; Input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain generated virtual demand classification samples and the corresponding low-code development data; Divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into a low-code small sample model, and obtain a trained low-code small sample model; Inputting new demand data into the trained low-code small sample model to obtain output low-code development data; The virtual demand classification samples and the corresponding low-code development data are divided into a training set and a test set, and the training set is input into a low-code small sample model to obtain a trained low-code small sample model, wherein the low-code small sample model includes a YOLOv3 model, including: Input the training set into the feature convolution module to obtain output first dimension data; Processing the first dimensional data through a residual module to obtain second dimensional data; The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data; Input the third dimension data into the concat operation module to obtain three feature data; The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, and the weight parameters of the support vector machine model module are adjusted to obtain a trained low-code small sample model; The step of inputting the training set into a feature convolution module to obtain output first dimension data includes: The image data and text data of the training set are identified; Extracting text content of the image data and converting it into first feature sequence data; Converting the text data into second feature sequence data; The first feature sequence data and the second feature sequence data are combined into feature sequence data, and the feature sequence data is input into a feature convolution module to obtain output first dimensional data.

2. The application code development method based on the artificial intelligence model according to claim 1 is characterized in that: The network demand data is classified to obtain demand classification samples of different dimensions, including: The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

3. The application code development method based on the artificial intelligence model according to claim 2 is characterized in that: The step of extracting the demand features of the network demand data and classifying the network demand data according to the demand features to obtain demand classification samples includes: Extracting industrial dimension features, agricultural dimension features, and service industry dimension features of the network demand data; The network demand data is classified according to the industrial dimension characteristics, agricultural dimension characteristics, and service industry dimension characteristics to obtain industrial demand samples, agricultural demand samples, and service industry demand samples.

4. The method for developing application code based on an artificial intelligence model according to claim 3, characterized in that: The step of inputting the demand classification sample and the corresponding low-code development data into a conditional generative adversarial network to obtain a generated virtual demand classification sample and the corresponding low-code development data includes: Calculate the quantitative proportion of the industrial demand samples, agricultural demand samples, and service industry demand samples; The number and proportion of industrial demand samples, agricultural demand samples, and service industry demand samples output by the model remain unchanged and are set as the conditional information of the generator network in the conditional generative adversarial network; The generator network receives random noise and conditional information as input to generate virtual data that matches the conditions. The discriminator network receives industrial demand samples, agricultural demand samples, service industry demand samples and corresponding low-code development data and conditional information to determine the authenticity of the virtual data, and obtains output virtual demand classification samples and corresponding low-code development data.

5. An application code development system based on an artificial intelligence model, characterized in that: The system comprises: A collection module, used to collect network demand data; A demand classification module, used to classify the network demand data to obtain demand classification samples of different dimensions; A low-code development data module is used to obtain low-code development data according to the required classification samples; A generation module, used to input the demand classification samples and the corresponding low-code development data into a conditional generative adversarial network to obtain a generated virtual demand classification sample and the corresponding low-code development data; A training module, used to divide the virtual demand classification samples and the corresponding low-code development data into a training set and a test set, input the training set into the low-code small sample model, and obtain a trained low-code small sample model; An output module, used to input new demand data into the trained low-code small sample model to obtain output low-code development data; The virtual demand classification samples and the corresponding low-code development data are divided into a training set and a test set, and the training set is input into a low-code small sample model to obtain a trained low-code small sample model, wherein the low-code small sample model includes a YOLOv3 model, including: Input the training set into the feature convolution module to obtain output first dimension data; Processing the first dimensional data through a residual module to obtain second dimensional data; The second dimensional data is enlarged by an upper pooling module to obtain third dimensional data; Input the third dimension data into the concat operation module to obtain three feature data; The three feature data are input into the support vector machine model module, updated iteratively, the loss function of the support vector machine model module is optimized, and the weight parameters of the support vector machine model module are adjusted to obtain a trained low-code small sample model; The step of inputting the training set into a feature convolution module to obtain output first dimension data includes: The image data and text data of the training set are identified; Extracting text content of the image data and converting it into first feature sequence data; Converting the text data into second feature sequence data; The first feature sequence data and the second feature sequence data are combined into feature sequence data, and the feature sequence data is input into a feature convolution module to obtain output first dimensional data.

6. The application code development system based on artificial intelligence model according to claim 5, characterized in that: The network demand data is classified to obtain demand classification samples of different dimensions, including: The demand features of the network demand data are extracted, and the network demand data are classified according to the demand features to obtain demand classification samples.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the application code development method based on the artificial intelligence model described in any one of claims 1 to 4 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the application code development method based on the artificial intelligence model described in any one of claims 1 to 4 are implemented.

Images