Government affair platform information data security processing system and method

Through natural language processing technology based on deep learning, multi-dimensional semantic understanding of government platform information data is solved, and the shortcomings of traditional methods are solved when processing complex information data are realized, intelligent judgment and corresponding protection measures are achieved for the level of information confidentiality, and the improvement of the Security of government data.

CN119089459AInactive Publication Date: 2024-12-06FUJIAN ZHONGXIN NET SAFETY INFORMATION TECHNOLOGY CO LTD

Patent Information

Application Number
CN202411068649.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-12-06
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional information security processing methods have limited processing capabilities for complex information data, which are prone to misjudgments and are difficult to effectively protect sensitive information in government affairs platforms.

Method used

The natural language processing technology based on deep learning is adopted to carry out multi-dimensional semantic understanding of government platform information data, and through multi-scale semantic correlation coding and bidirectional loop correlation coding, semantic features in the data are captured, and the confidentiality level of information is intelligently judged.

Benefits of technology

It improves the security of information data on government affairs platforms, and can take corresponding confidentiality measures according to different information confidentiality levels to enhance data protection capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089459A_ABST
    Figure CN119089459A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of information security, and particularly discloses a government affair platform information data security processing system and method.The government affair platform information data to be processed are subjected to multi-dimensional semantic understanding by adopting a natural language processing technology based on deep learning, semantic feature expressions in the government affair platform information data are mined, and the semantic feature expressions in the government affair platform information data are extracted; the sensitive information in the data is captured, so that the confidentiality level of the information is intelligently judged. Therefore, different information secrecy measures can be adopted based on different information secrecy levels, so that the security of government affair platform information data is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of information security technology, and more specifically, to a system and method for securely processing information data on a government affairs platform. Background Art

[0002] With the rapid development of information technology, the security processing of information data on government platforms has become an important issue. Government platforms usually contain a large amount of information data from various fields of public departments and related institutions and personal information of citizens, such as administrative regulations, economic statistics, citizen identity information, financial data, health records, etc. The leakage or hacking of these data may have a serious impact on national security, citizen privacy and social stability.

[0003] In order to ensure that sensitive information in the government platform is not accessed, leaked or tampered by unauthorized personnel, different information confidentiality measures need to be taken according to the sensitivity of the information data of the government platform. However, traditional information security processing methods usually use simple keyword filtering, which has limited processing capabilities for complex information data and is prone to misjudgment. Therefore, in order to improve the security processing capabilities of government platform information data, an optimized government platform information data security processing system and method is expected. Summary of the invention

[0004] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides a system and method for secure processing of government platform information data, which uses natural language processing technology based on deep learning to perform multi-dimensional semantic understanding of the government platform information data to be processed, and mines the semantic feature expressions in the government platform information data to capture sensitive information in the data, thereby intelligently judging the confidentiality level of the information. In this way, different information confidentiality measures can be adopted based on different information confidentiality levels to improve the security of government platform information data.

[0005] Accordingly, according to one aspect of the present application, a method for securely processing information data on an e-government platform is provided, comprising:

[0006] Obtain information data on government affairs platforms to be processed;

[0007] Perform multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector;

[0008] Fusion of the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector;

[0009] Based on the multi-dimensional government affairs platform information semantic understanding feature vector, the information confidentiality level type is determined.

[0010] In the above-mentioned government affairs platform information data security processing method, multi-scale semantic understanding is performed on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector, including: performing word segmentation processing on the government affairs platform information data to be processed and then passing it through a word embedding layer to obtain a sequence of government affairs platform information word embedding coding feature vectors; performing local semantic association coding on the sequence of government affairs platform information word embedding coding feature vectors to obtain the first government affairs platform information semantic understanding feature vector; performing bidirectional cyclic association coding on the sequence of government affairs platform information word embedding coding feature vectors to obtain the second government affairs platform information semantic understanding feature vector.

[0011] In the above-mentioned government platform information data security processing method, the sequence of the government platform information word embedding encoding feature vectors is locally semantically associated encoded to obtain the first government platform information semantic understanding feature vector, including: passing the sequence of the government platform information word embedding encoding feature vectors through the first government platform information semantic encoder based on the convolutional neural network model to obtain the first government platform information semantic understanding feature vector.

[0012] In the above-mentioned government platform information data security processing method, the sequence of the government platform information word embedding encoding feature vectors is bidirectionally cyclically associated encoded to obtain the second government platform information semantic understanding feature vector, including: passing the sequence of the government platform information word embedding encoding feature vectors through a second government platform information semantic encoder based on a bidirectional gated recurrent unit to obtain the second government platform information semantic understanding feature vector.

[0013] In the above-mentioned government affairs platform information data security processing method, the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector are fused to obtain a multi-dimensional government affairs platform information semantic understanding feature vector, including: using a projection layer to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector.

[0014] In the above-mentioned government affairs platform information data security processing method, a projection layer is used to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector, including: using the following projection feature fusion formula to perform feature fusion on the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector; wherein the projection feature fusion formula is:

[0015]

[0016] Among them, V1 is the semantic understanding feature vector of the first government platform information, V2 is the semantic understanding feature vector of the second government platform information, V f is the semantic understanding feature vector of the multi-dimensional government affairs platform information, represents projection fusion processing, and [·;·] represents cascade processing.

[0017] In the above-mentioned government platform information data security processing method, the information confidentiality level type is determined based on the multi-dimensional government platform information semantic understanding feature vector, including: passing the multi-dimensional government platform information semantic understanding feature vector through a classifier to obtain a classification result, and the classification result is used to represent the information confidentiality level type label.

[0018] The above-mentioned government platform information data security processing method also includes a training step: training the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

[0019] In the above-mentioned government platform information data security processing method, the training step includes: obtaining training data, the training data includes training government platform information data to be processed, and the true value of the information confidentiality level type; performing word segmentation on the training government platform information data to be processed and passing it through the word embedding layer to obtain a sequence of training government platform information word embedding encoding feature vectors; passing the sequence of training government platform information word embedding encoding feature vectors through the first government platform information semantic encoder based on the convolutional neural network model to obtain a training first government platform information semantic understanding feature vector; passing the sequence of training government platform information word embedding encoding feature vectors through the second government platform information semantic encoder based on the bidirectional gated recurrent unit to obtain a training second government platform information semantic understanding feature vector; using A projection layer is used to fuse the trained first government information platform information semantic understanding feature vector and the trained second government information platform information semantic understanding feature vector to obtain a trained multi-dimensional government information platform information semantic understanding feature vector; the trained multi-dimensional government information platform information semantic understanding feature vector is passed through the classifier to obtain a classification loss function value; based on the trained multi-dimensional government information platform information semantic understanding feature vector, a multi-dimensional government information semantic understanding loss function value is calculated; and a weighted sum of the classification loss function value and the multi-dimensional government information semantic understanding loss function value is used as the loss function value to train the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

[0020] According to another aspect of the present application, a government affairs platform information data security processing system is provided, which includes:

[0021] The government affairs platform information data acquisition module is used to acquire the government affairs platform information data to be processed;

[0022] A multi-scale semantic understanding module, used for performing multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector;

[0023] A multi-dimensional semantic feature fusion module, used to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector;

[0024] The confidentiality level type generation module is used to determine the information confidentiality level type based on the multi-dimensional government affairs platform information semantic understanding feature vector.

[0025] Compared with the prior art, the government platform information data security processing system and method provided by this application uses natural language processing technology based on deep learning to perform multi-dimensional semantic understanding on the government platform information data to be processed, and mines the semantic feature expressions in the government platform information data to capture sensitive information in the data, thereby intelligently judging the confidentiality level of the information. In this way, different information confidentiality measures can be adopted based on different information confidentiality levels to improve the security of government platform information data. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0027] Figure 1 It is a flowchart of a method for securely processing information data of a government affairs platform according to an embodiment of the present application.

[0028] Figure 2 It is a schematic diagram of the architecture of the method for securely processing information data of the government affairs platform according to the embodiment of the present application.

[0029] Figure 3 This is a flowchart of performing multi-scale semantic understanding on the government affairs platform information data to be processed in the method for securely processing government affairs platform information data according to an embodiment of the present application to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector.

[0030] Figure 4 It is a flowchart of the training steps in the method for securely processing information data of the government affairs platform according to an embodiment of the present application.

[0031] Figure 5 It is a block diagram of the government affairs platform information data security processing system according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] Below, the embodiments of the present application will be described in more detail in conjunction with the accompanying drawings, and the above-mentioned and other purposes, features and advantages of the present application will become more obvious. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein. At the same time, the accompanying drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used together with the embodiments of the present application to explain the present application, and do not constitute a limitation on the present application. In the drawings, the same reference numerals generally represent the same parts or steps.

[0033] Figure 1 It is a flowchart of a method for securely processing information data of a government affairs platform according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of the method for securely processing information data on the government affairs platform according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the method for securely processing government platform information data according to the embodiment of the present application includes the steps of: S110, obtaining the government platform information data to be processed; S120, performing multi-scale semantic understanding on the government platform information data to be processed to obtain a first government platform information semantic understanding feature vector and a second government platform information semantic understanding feature vector; S130, fusing the first government platform information semantic understanding feature vector and the second government platform information semantic understanding feature vector to obtain a multi-dimensional government platform information semantic understanding feature vector; S140, determining the information confidentiality level type based on the multi-dimensional government platform information semantic understanding feature vector.

[0034] As mentioned in the background technology above, in order to ensure that the information data of the government platform is not accessed, leaked or tampered by unauthorized personnel, the information in the government platform can be classified and marked according to the sensitivity level of the information, such as public information, internal information and confidential information. In this way, by marking the sensitivity level of the information, corresponding confidentiality measures can be taken in the subsequent processing and storage process. However, traditional information security processing methods usually use simple keyword filtering, which has limited processing capabilities for complex information data and is prone to misjudgment.

[0035] In response to the above technical problems, the technical concept of this application is to use natural language processing technology based on deep learning to perform multi-dimensional semantic understanding on the government platform information data to be processed, and to mine the semantic feature expressions in the government platform information data to capture sensitive information in the data, so as to intelligently judge the confidentiality level of the information. In this way, different information confidentiality measures can be adopted based on different information confidentiality levels to improve the security of government platform information data.

[0036] In the above-mentioned government platform information data security processing method, the step S110 obtains the government platform information data to be processed. It should be understood that the government platform information data refers to various types of government data generated, transmitted and stored on the government platform, including government documents, announcements, administrative decisions, statistical data, etc. These government platform information data usually have different levels of sensitivity, such as personal privacy, commercial secrets, economic statistics, etc. Further using deep learning technology to perform semantic analysis and understanding on these government platform information data, it is possible to extract the semantic feature expressions therein, thereby identifying the sensitivity of the data based on the semantic features of the data, and judging the confidentiality level of the information, so as to achieve the confidentiality management task of the government platform information.

[0037] In the above-mentioned government affairs platform information data security processing method, the step S120 performs multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector. Specifically, Figure 3 This is a flowchart of performing multi-scale semantic understanding on the government platform information data to be processed in the government platform information data security processing method according to the embodiment of the present application to obtain the first government platform information semantic understanding feature vector and the second government platform information semantic understanding feature vector. Figure 3 As shown, the step S120 includes: S121, performing word segmentation on the government affairs platform information data to be processed and then passing it through a word embedding layer to obtain a sequence of government affairs platform information word embedding coding feature vectors; S122, performing local semantic association coding on the sequence of government affairs platform information word embedding coding feature vectors to obtain the first government affairs platform information semantic understanding feature vector; S123, performing bidirectional cyclic association coding on the sequence of government affairs platform information word embedding coding feature vectors to obtain the second government affairs platform information semantic understanding feature vector.

[0038] Specifically, in step S121, the government affairs platform information data to be processed is processed by word segmentation and then passed through a word embedding layer to obtain a sequence of government affairs platform information word embedding encoding feature vectors. It should be understood that the word embedding layer is a commonly used technology in deep learning, which is used to represent discrete words as vectors in a continuous vector space. The word embedding layer can map each word to a real number vector of a fixed length, so that the semantic relationship between words can be reflected in the vector space. In the technical solution of the present application, the government affairs platform information data to be processed is first processed by word segmentation to divide the continuous text sequence into discrete word units, and then each word is mapped to a high-dimensional vector representation of a fixed length through word embedding coding to facilitate subsequent data processing and calculation. In addition, the word embedding layer can capture the semantic relationship between words by learning the distributed representation of words in a high-dimensional feature space, so that the computer can understand and process the meaning in the government affairs platform information data to be processed, and provide a basis for semantic understanding and feature representation for subsequent data analysis tasks.

[0039] Specifically, in step S122, the sequence of the word embedding encoding feature vectors of the government platform information is locally semantically associated encoded to obtain the first government platform information semantic understanding feature vector. In a specific example of the present application, the implementation method of locally semantically associated encoding the sequence of the word embedding encoding feature vectors of the government platform information to obtain the first government platform information semantic understanding feature vector is to pass the sequence of the word embedding encoding feature vectors of the government platform information through the first government platform information semantic encoder based on the convolutional neural network model to obtain the first government platform information semantic understanding feature vector. A person of ordinary skill in the art should know that a convolutional neural network (CNN) is a powerful feature extractor that can automatically learn and extract local features from input data. In the technical solution of the present application, the sequence of the word embedding encoding feature vectors of the government platform information is input into the first government platform information semantic encoder based on the convolutional neural network model, and the first government platform information semantic encoder performs a local convolution operation on the sequence of the word embedding encoding feature vectors of the government platform information by applying a filter of a convolutional layer, so as to capture the semantic features of different positions in the government platform information, such as keywords, phrases, and sentence structures, so as to better understand the meaning of the government platform information data. In addition, when performing convolution operations, the first government platform information semantic encoder based on the convolutional neural network model can consider the local contextual relationship between adjacent words in the input sequence through a sliding window, so that the convolutional neural network can capture the semantic correlation between short-distance words, such as the co-occurrence relationship between words, the dependency relationship between words, etc., thereby providing a more comprehensive semantic understanding of the government platform information data and providing more accurate and comprehensive feature representation for subsequent information confidentiality level judgment tasks.

[0040] Specifically, in step S123, the sequence of the word embedding encoding feature vectors of the government platform information is subjected to bidirectional cyclic association encoding to obtain the second government platform information semantic understanding feature vector. In a specific example of the present application, the sequence of the word embedding encoding feature vectors of the government platform information is subjected to bidirectional cyclic association encoding to obtain the second government platform information semantic understanding feature vector by passing the sequence of the word embedding encoding feature vectors of the government platform information through a second government platform information semantic encoder based on a bidirectional gated recurrent unit to obtain the second government platform information semantic understanding feature vector. It should be understood that the bidirectional gated recurrent unit (Bidirectional Gated Recurrent Unit, BiGRU) is a recurrent neural network capable of modeling contextual information. In traditional recurrent neural networks, information is transmitted unidirectionally from the past to the future in time, and only the contextual information before the current moment can be considered. The bidirectional gated recurrent unit realizes the transmission and modeling of bidirectional information by running two independent gated recurrent units (Gated Recurrent Unit, GRU) in the forward and reverse directions, so that the network can capture the contextual information of the past and the future at the same time. In the technical solution of the present application, the meaning of words and contextual relationships in the government platform information data often depend on the context of the entire sentence or paragraph. By using a second government platform information semantic encoder based on a bidirectional gated recurrent unit to perform cyclic association encoding on the sequence of the government platform information word embedding encoding feature vectors, the semantic context features in the government platform information can be better modeled, providing richer and more accurate semantic understanding. In addition, there are usually long-distance dependencies in the government platform information, and it is necessary to consider the more distant contextual information in order to understand the semantic information therein. The bidirectional gated recurrent unit has a good long-term dependency modeling capability, and can control the flow of information through a gating mechanism, helping the model to better capture the long-term dependencies in the information, thereby improving the model's semantic understanding of government platform information.

[0041] In the above-mentioned government affairs platform information data security processing method, the step S130 is to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector. In a specific example of the present application, the implementation method of fusing the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector is to use a projection layer to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector. It should be understood that the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector respectively capture the semantic feature representation of government affairs platform information at different scales. Through the fusion processing of the projection layer, the semantic information of the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector can be complemented and enriched to obtain a more comprehensive multi-dimensional government affairs platform information semantic understanding feature vector. In addition, since the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector come from different encoders, they exist in different feature representation spaces. Therefore, in order to effectively integrate the two, dimensional matching and alignment are required. The projection layer can map the first government information platform semantic understanding feature vector and the second government information platform semantic understanding feature vector to the same dimensional space, and establish the semantic association between the first government information platform semantic understanding feature vector and the second government information platform semantic understanding feature vector by learning the projection matrix, so as to capture the correlation and association between contextual semantic features of different scales, so as to effectively integrate the first government information platform semantic understanding feature vector and the second government information platform semantic understanding feature vector, and further improve the semantic understanding and semantic expression capabilities of government information.

[0042] Specifically, using a projection layer to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector, including: using the following projection feature fusion formula to perform feature fusion on the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector; wherein the projection feature fusion formula is:

[0043]

[0044] Among them, V1 is the semantic understanding feature vector of the first government platform information, V2 is the semantic understanding feature vector of the second government platform information, Vf is the semantic understanding feature vector of the multi-dimensional government affairs platform information, represents projection fusion processing, and [·;·] represents cascade processing.

[0045] In the above-mentioned government affairs platform information data security processing method, the step S140 determines the type of information confidentiality level based on the multi-dimensional government affairs platform information semantic understanding feature vector. In a specific example of the present application, the implementation method of determining the type of information confidentiality level based on the multi-dimensional government affairs platform information semantic understanding feature vector is to pass the multi-dimensional government affairs platform information semantic understanding feature vector through a classifier to obtain a classification result, and the classification result is used to represent the type label of information confidentiality level. It should be understood that the information in the government affairs platform usually contains various types and different confidentiality levels. By inputting the multi-dimensional government affairs platform information semantic understanding feature vector into the classifier, the information can be automatically classified using a machine learning algorithm and divided into different confidentiality level types. That is, by training the classifier, it learns the association between the multi-dimensional government affairs platform information semantic understanding feature vector and different confidentiality level types, thereby realizing the classification of information, so as to realize the automatic classification and labeling of government affairs platform information. In this way, the government affairs platform information can be labeled and managed according to its confidentiality level to ensure the security and confidentiality of the information.

[0046] It should be understood that before using the above-mentioned neural network model, it is necessary to train the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier. That is to say, in the government platform information data security processing method of the present application, it also includes a training step of training the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

[0047] Figure 4 FIG. 1 is a flowchart of the training steps in the method for securely processing information data of the government affairs platform according to an embodiment of the present application. Figure 4As shown, the training step includes: S210, obtaining training data, the training data including training government affairs platform information data to be processed, and the true value of the information confidentiality level type; S220, performing word segmentation on the training government affairs platform information data to be processed and passing it through the word embedding layer to obtain a sequence of training government affairs platform information word embedding encoding feature vectors; S230, passing the sequence of training government affairs platform information word embedding encoding feature vectors through the first government platform information semantic encoder based on the convolutional neural network model to obtain a training first government affairs platform information semantic understanding feature vector; S240, passing the sequence of training government affairs platform information word embedding encoding feature vectors through the second government platform information semantic encoder based on the bidirectional gated recurrent unit to obtain a training second government affairs platform information semantic understanding feature vector; S250, using the projection layer To fuse the trained first government affairs platform information semantic understanding feature vector and the trained second government affairs platform information semantic understanding feature vector to obtain a trained multi-dimensional government affairs platform information semantic understanding feature vector; S260, pass the trained multi-dimensional government affairs platform information semantic understanding feature vector through the classifier to obtain a classification loss function value; S270, based on the trained multi-dimensional government affairs platform information semantic understanding feature vector, calculate the multi-dimensional government affairs platform information semantic understanding loss function value; S280, use the weighted sum of the classification loss function value and the multi-dimensional government affairs platform information semantic understanding loss function value as the loss function value to train the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

[0048] In the above technical solution, the first training government affairs platform information semantic understanding feature vector and the second training government affairs platform information semantic understanding feature vector respectively express the local convolution scale semantic association features based on word segmentation semantics and the bidirectional gated scale cyclic association features of the training government affairs platform information data to be processed. Therefore, when the projection layer is used to fuse the first training government affairs platform information semantic understanding feature vector and the second training government affairs platform information semantic understanding feature vector, considering the difference in text semantic feature extraction modes based on word segmentation semantics corresponding to the first training government affairs platform information semantic understanding feature vector and the second training government affairs platform information semantic understanding feature vector, the first training government affairs platform information semantic understanding feature vector and the second training government affairs platform information semantic understanding feature vector will have different feature group density representations based on semantic distribution patterns, which will lead to the complication of the feature distribution representation structure based on different feature group density representations of the training multi-dimensional government affairs platform information semantic understanding feature vector when it is fused, and there is a relatively obvious problem of insufficient classification understanding of complex feature expressions, which affects the accuracy of the classification results obtained by the classifier of the training multi-dimensional government affairs platform information semantic understanding feature vector.

[0049] Based on this, the present application further introduces a loss function for promoting classification understanding of the complex feature expression of the training multi-dimensional government affairs platform information semantic understanding feature vector, that is, based on the training multi-dimensional government affairs platform information semantic understanding feature vector, the multi-dimensional government affairs platform information semantic understanding loss function value is calculated. Among them, based on the training multi-dimensional government affairs platform information semantic understanding feature vector, the multi-dimensional government affairs platform information semantic understanding loss function value is calculated, including: first, by calculating the mean and variance between the eigenvalues ​​of any two positions of the training multi-dimensional government affairs platform information semantic understanding feature vector to obtain the mean weight matrix and the variance weight matrix, and then after query-multiplying the training multi-dimensional government affairs platform information semantic understanding feature vector with the mean weight matrix and the variance weight matrix respectively, the Frobenius norm of the autocorrelation matrix of the two eigenvectors is calculated, and the product of the Frobenius norm of the autocorrelation matrix of the training multi-dimensional government affairs platform information semantic understanding feature vector and the weight as a hyperparameter is subtracted to obtain the multi-dimensional government affairs platform information semantic understanding loss function value.

[0050] For example, expressed as

[0051]

[0052] Where V represents the semantic understanding feature vector of the training multi-dimensional government affairs platform information, M μ is the mean weight matrix, M σ is the variance weight matrix, T represents the transpose of the vector, represents the matrix multiplication operation, α is the weight as a hyperparameter, ‖·‖ F represents the Frobenius norm of the matrix, Loss represents the loss function value of the information semantic understanding of the multi-dimensional government affairs platform, and as described above:

[0053]

[0054] Among them, v i represents the i-th eigenvalue of the training multi-dimensional government affairs platform information semantic understanding feature vector, v j represents the jth eigenvalue of the training multi-dimensional government affairs platform information semantic understanding feature vector, M μ (i,j) represents the eigenvalue of the (i,j)th position of the mean weight matrix, M σ (i,j) represents the eigenvalue at the (i,j)th position of the variance weight matrix.

[0055] That is, the multi-dimensional government platform information semantic understanding loss function value is based on determining the overall distributed structural bottleneck of the training multi-dimensional government platform information semantic understanding feature vector through the multi-dimensional interaction of the detail group of the training multi-dimensional government platform information semantic understanding feature vector, and simulates the correlation between the structural details and the macro-structured feature representation behavior of the training multi-dimensional government platform information semantic understanding feature vector through the low-rank structured reasoning of the training multi-dimensional government platform information semantic understanding feature vector. In this way, by training the classifier with the multi-dimensional government platform information semantic understanding loss function value, the structural detail dependency relative to the macro-complexity can be modeled in the classification scenario, thereby promoting the understanding of the macro-complex feature expression based on the structural details of the training multi-dimensional government platform information semantic understanding feature vector, and improving the accuracy of the classification results.

[0056] In summary, the method for processing government platform information data security according to the embodiment of the present application is explained, which uses natural language processing technology based on deep learning to perform multi-dimensional semantic understanding on the government platform information data to be processed, and mines the semantic feature expressions in the government platform information data to capture sensitive information in the data, thereby intelligently judging the confidentiality level of the information. In this way, different information confidentiality measures can be adopted based on different information confidentiality levels to improve the security of government platform information data.

[0057] Figure 5 FIG. 1 is a block diagram of a government affairs platform information data security processing system according to an embodiment of the present application. Figure 5As shown, according to the embodiment of the present application, the government affairs platform information data security processing system 100 includes: a government affairs platform information data acquisition module 110, used to acquire the government affairs platform information data to be processed; a multi-scale semantic understanding module 120, used to perform multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector; a multi-dimensional semantic feature fusion module 130, used to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector; a confidentiality level type generation module 140, used to determine the information confidentiality level type based on the multi-dimensional government affairs platform information semantic understanding feature vector.

[0058] Here, those skilled in the art can understand that the specific operations of each module in the above-mentioned government platform information data security processing system have been referenced above. Figures 1 to 4 The method for securely processing information data on the government platform has been introduced in detail, and therefore, its repeated description will be omitted.

[0059] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0060] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0061] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0062] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for securely processing information data on a government platform, characterized in that: include: Obtain information data on government affairs platforms to be processed; Perform multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector; Fusion of the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector; Based on the multi-dimensional government affairs platform information semantic understanding feature vector, the information confidentiality level type is determined.

2. The method for securely processing information data on a government affairs platform according to claim 1 is characterized in that: Performing multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector, including: After word segmentation processing is performed on the government affairs platform information data to be processed, the data is passed through a word embedding layer to obtain a sequence of government affairs platform information word embedding encoding feature vectors; Performing local semantic association coding on the sequence of word embedding coding feature vectors of the government affairs platform information to obtain the semantic understanding feature vector of the first government affairs platform information; The sequence of the word embedding encoding feature vectors of the government affairs platform information is subjected to bidirectional cyclic association encoding to obtain the second government affairs platform information semantic understanding feature vector.

3. The method for securely processing information data on the government affairs platform according to claim 2 is characterized in that: Performing local semantic association coding on the sequence of the word embedding coding feature vectors of the government affairs platform information to obtain the semantic understanding feature vector of the first government affairs platform information includes: The sequence of the government platform information word embedding encoding feature vectors is passed through the first government platform information semantic encoder based on a convolutional neural network model to obtain the first government platform information semantic understanding feature vector.

4. The method for securely processing information data on the government affairs platform according to claim 3 is characterized in that: Performing bidirectional cyclic association coding on the sequence of the word embedding coding feature vectors of the government affairs platform information to obtain the second government affairs platform information semantic understanding feature vector, including: The sequence of the government affairs platform information word embedding encoding feature vectors is passed through a second government affairs platform information semantic encoder based on a bidirectional gated recurrent unit to obtain the second government affairs platform information semantic understanding feature vector.

5. The method for securely processing information data on a government affairs platform according to claim 4 is characterized in that: The first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector are integrated to obtain a multi-dimensional government affairs platform information semantic understanding feature vector, including: A projection layer is used to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector.

6. The method for securely processing information data on the government affairs platform according to claim 5 is characterized in that: Using a projection layer to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain the multi-dimensional government affairs platform information semantic understanding feature vector, including: The first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector are feature-fused using the following projection feature fusion formula to obtain the multi-dimensional government affairs platform information semantic understanding feature vector; wherein the projection feature fusion formula is: Among them, V1 is the semantic understanding feature vector of the first government platform information, V2 is the semantic understanding feature vector of the second government platform information, V f is the semantic understanding feature vector of the multi-dimensional government affairs platform information, represents projection fusion processing, and [·;·] represents cascade processing.

7. The method for securely processing information data on the government affairs platform according to claim 6 is characterized in that: Based on the multi-dimensional government affairs platform information semantic understanding feature vector, the type of information confidentiality level is determined, including: The multi-dimensional government affairs platform information semantic understanding feature vector is passed through a classifier to obtain a classification result, and the classification result is used to represent the information confidentiality level type label.

8. The method for securely processing information data on a government affairs platform according to claim 7 is characterized in that: It also includes a training step: training the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

9. The method for securely processing information data on a government affairs platform according to claim 8 is characterized in that: The training step comprises: Acquire training data, wherein the training data includes training government affairs platform information data to be processed and true values ​​of information confidentiality level types; After word segmentation processing is performed on the training government affairs platform information data to be processed, the data is passed through the word embedding layer to obtain a sequence of training government affairs platform information word embedding encoding feature vectors; Passing the sequence of the training government affairs platform information word embedding encoding feature vectors through the first government affairs platform information semantic encoder based on the convolutional neural network model to obtain the training first government affairs platform information semantic understanding feature vector; Passing the sequence of the training government affairs platform information word embedding encoding feature vectors through the second government affairs platform information semantic encoder based on the bidirectional gated recurrent unit to obtain the training second government affairs platform information semantic understanding feature vector; Using a projection layer to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector; Passing the training multi-dimensional government affairs platform information semantic understanding feature vector through the classifier to obtain a classification loss function value; Calculating the multi-dimensional government affairs platform information semantic understanding loss function value based on the training multi-dimensional government affairs platform information semantic understanding feature vector; The weighted sum of the classification loss function value and the multi-dimensional government platform information semantic understanding loss function value is used as the loss function value to train the word embedding layer, the first government platform information semantic encoder based on the convolutional neural network model, the second government platform information semantic encoder based on the bidirectional gated recurrent unit, the projection layer and the classifier.

10. A government affairs platform information data security processing system, characterized in that: include: The government affairs platform information data acquisition module is used to acquire the government affairs platform information data to be processed; A multi-scale semantic understanding module, used for performing multi-scale semantic understanding on the government affairs platform information data to be processed to obtain a first government affairs platform information semantic understanding feature vector and a second government affairs platform information semantic understanding feature vector; A multi-dimensional semantic feature fusion module, used to fuse the first government affairs platform information semantic understanding feature vector and the second government affairs platform information semantic understanding feature vector to obtain a multi-dimensional government affairs platform information semantic understanding feature vector; The confidentiality level type generation module is used to determine the information confidentiality level type based on the multi-dimensional government affairs platform information semantic understanding feature vector.

Citation Information

Patent Citations

  • Short text classification method based on semantic enhancement and multi-level label embedding

    CN110866117A

  • Method and device for automatically determining secret grade of secret-related text

    CN112347779A

  • Government affair sensitive information identification method and device, equipment, medium and program product

    CN114386408A

  • Government information management system and method based on intelligent AI

    CN117349437A

  • Multi-scale feature fusion news theme classification model construction method, news theme classification method and related device

    CN118193740A

Cited By

  • Standardized management system for aviation enterprises

    CN119850149A