Information desensitization method and system, electronic equipment and storage medium

Abstract

The invention discloses an information desensitization method and system, electronic equipment and a storage medium, and the method comprises the steps: firstly receiving a first query request of a user, recognizing each piece of target information in the query request according to a recognition rule, converting each piece of target information into a set identifier, and obtaining a second query request containing the set identifier; and calling an API key corresponding to the LLM model, and sending the second query request and the API key to a server providing the LLM model service. In this way, secret-related and / or privacy information in the query request of the user can be effectively filtered and hidden, and leakage of the secret-related and / or privacy information in the query request is avoided.

Description

Technical Field

[0001] The present application relates to the field of information security technology, and in particular to an information desensitization method, system, electronic device and storage medium. Background Art

[0002] With the widespread application of large language models (LLMs) in various industries, the huge potential they demonstrate is changing the way of working in various fields. However, the abuse of these models is also gradually increasing. Criminals use technical means such as prompt injection to break through the inherent security limitations of LLMs, resulting in increasing security threats. These threats include illegal activities such as writing malicious scripts to conduct network attacks and creating phishing emails through automated means.

[0003] Although using the cloud services of LLM service providers can avoid some hardware requirements, it also raises serious concerns about data privacy and security. In particular, when government and enterprise developers use LLM to generate code, they may inadvertently leak confidential information such as keys and IP addresses. This risk causes private information and confidential information to be leaked when government and enterprises apply LLM. Summary of the invention

[0004] The present application provides an information desensitization method, system, electronic device and storage medium to avoid leakage of private information and confidential information during information query.

[0005] In a first aspect, an embodiment of the present application provides an information desensitization method, the method comprising:

[0006] Receive a first query request from a user, wherein the first query request at least includes a request content and a target large language model LLM for analyzing the request content;

[0007] According to the identification rule, each target information is identified in the first query request, wherein the target information is confidential information and / or private information;

[0008] Convert each of the target information into a setting identifier to obtain a second query request including the setting identifier;

[0009] Retrieve the application programming interface API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service.

[0010] Through the above method, the target information can be identified in the query request, and then the target information is replaced with the set identifier, and the query request containing the set identifier is sent to the server providing LLM, thereby ensuring the security of the data.

[0011] In an optional embodiment, after converting each of the target information into a setting identifier, the method further includes:

[0012] Create a mapping relationship between target information and corresponding setting identifiers;

[0013] The corresponding relationship is saved in the target database.

[0014] By using the above method, a mapping relationship between confidential information and set identifiers is established, which can further improve the security and integrity of the information.

[0015] In an optional embodiment, after obtaining the second query request including the set identifier, the method further includes:

[0016] Inferring the set identifier in the second query request through the inference model to obtain inference information corresponding to the set identifier;

[0017] Determining whether the inference information is consistent with the target information corresponding to the setting identifier in the mapping relationship;

[0018] If they are consistent, the desensitization rules for desensitizing the query request are updated according to the inference result;

[0019] If there is no consistency, the desensitization rule is maintained.

[0020] In this method, a discrimination component can also be used to confirm whether the desensitization process is effective, thereby further improving the security of confidential information.

[0021] In an optional embodiment, creating a correspondence between target information and a corresponding setting identifier includes:

[0022] Retrieving an identifier database containing various types of setting identifiers;

[0023] A corresponding setting identifier is allocated to each target information, and a mapping relationship between the target information and the setting identifier is established.

[0024] In an optional embodiment, retrieving an application programming interface API key corresponding to the LLM model includes:

[0025] Determine the available API keys in the API key collection;

[0026] Detecting whether there is an API key corresponding to the LLM model;

[0027] If it exists, retrieve the API key corresponding to the LLM model;

[0028] If it does not exist, retrieve any API key from the available API keys.

[0029] Through the above method, users can access LLM services through anonymous or virtual identities through API keys, further reducing the possibility of privacy information leakage.

[0030] In a second aspect, an embodiment of the present application provides an information desensitization system, the system comprising:

[0031] A receiving module, configured to receive a first query request from a user, wherein the first query request at least includes a request content and a target large language model LLM for analyzing the request content;

[0032] A processing module is used to identify each target information in the first query request according to the identification rule, convert each target information into a set identifier, and obtain a second query request containing the set identifier; retrieve the application programming interface API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service, wherein the target information is confidential information and / or privacy information.

[0033] In an optional embodiment, the processing module is further used to create a mapping relationship between the target information and the corresponding setting identifier; and save the corresponding relationship to the target database.

[0034] In an optional embodiment, the processing module is also used to infer the set identifier in the second query request through an inference model to obtain inference information corresponding to the set identifier; determine whether the inference information is consistent with the target information corresponding to the set identifier in the mapping relationship; if consistent, update the desensitization rules used to desensitize the query request according to the inference result; if inconsistent, maintain the desensitization rules.

[0035] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0036] Memory, used to store computer programs;

[0037] The processor is used to implement the steps of the above-mentioned information desensitization method when executing the computer program stored in the memory.

[0038] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned information desensitization method are implemented.

[0039] For each aspect from the second to the fourth aspect and the technical effects that may be achieved by each aspect, please refer to the above description of the technical effects that can be achieved by the first aspect or various possible schemes in the first aspect, and no further details will be given here. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of an information desensitization method provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of the structure of a data desensitization system provided in an embodiment of the present application;

[0042] Figure 3 A schematic diagram of the structure of the adversarial desensitization module provided in an embodiment of the present application;

[0043] Figure 4 A schematic diagram of the structure of the proxy scheduling module provided in the embodiment of the present application;

[0044] Figure 5 A workflow diagram for data query provided in an embodiment of the present application;

[0045] Figure 6 A schematic diagram of the structure of an information desensitization system provided in an embodiment of the present application;

[0046] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The specific operating methods in the method embodiments can also be applied to device embodiments or system embodiments. It should be noted that in the description of the present application, "multiple" is understood as "at least two". "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A is connected to B, which can represent: A is directly connected to B and A is connected to B through C. In addition, in the description of the present application, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0048] With the widespread application of large language models (LLMs) in various industries, the great potential they demonstrate is changing the way of working in various fields. However, the abuse of these models is also gradually increasing. Criminals use technical means such as prompt injection to break through the inherent security limitations of LLMs, resulting in increasing security threats. These threats include illegal activities such as writing malicious scripts in an automated way to carry out network attacks and creating phishing emails.

[0049] Although using the cloud services of LLM service providers can avoid some hardware requirements, it also raises serious concerns about data privacy and security. In particular, when government and enterprise developers use LLM to generate code, they may inadvertently leak confidential information such as keys and IP addresses. This risk causes private information and confidential information to be leaked when government and enterprises apply LLM.

[0050] In order to solve the above technical problems, the embodiment of the present application provides an information desensitization method, in which the first query request of the user is first received, and each target information is identified in the query request according to the identification rule, and each target information is converted into a set identifier to obtain a second query request containing the set identifier, and the application programming interface (Application Programming Interface, API) key corresponding to the LLM model is retrieved, and the second query request and the API key are sent to the server providing the LLM model service. In this way, the confidential and / or private information in the user's query request can be effectively filtered and hidden, avoiding the leakage of confidential and / or private information in the query request.

[0051] The technical solution of the present application is described in detail below through the accompanying drawings and specific embodiments.

[0052] Reference Figure 1 The figure is a flow chart of an information desensitization method provided in an embodiment of the present application, the method comprising:

[0053] S1, receiving the first query request from the user;

[0054] S2, identifying each target information in the first query request according to the identification rule;

[0055] S3, converting each target information into a set identifier to obtain a second query request including the set identifier;

[0056] S4, retrieve the API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service.

[0057] First of all, this method can be applied to a data desensitization system, refer to Figure 2 As shown, the data desensitization system includes an adversarial desensitization module and an agent scheduling module. In the data desensitization system, the adversarial desensitization module can be connected to the agent scheduling module.

[0058] Specifically, refer to Figure 3 As shown in FIG. 1 , the adversarial desensitization module includes a desensitization component, a discrimination component, and an anonymization log component. These three components may be interconnected.

[0059] When the adversarial desensitization module receives the user's first query request, it first identifies the information in the first query request through the desensitization component. The identification of the desensitization component includes named entity recognition and recognition rule recognition. First of all, the named entity recognition is to accurately identify the entity in the first query request through natural language processing technology. The entity can be personal or organization-related information such as name, work unit, address, etc.

[0060] In addition to identifying the entity in the first query request, the information in the first query request needs to be identified according to the identification rule. The identification rule can be an identification rule for sensitive word types pre-added by the user, that is, the identification rule can identify different sensitive words under the same type. Of course, different sensitive word types can be configured in the identification rule, so that different types of sensitive words can be identified through the identification rule. The sensitive word is confidential information or private information, such as IP address, API key, etc.

[0061] Therefore, the confidential information and / or private information in the first query request can be identified through the above-mentioned desensitization component. Then, the identified confidential information and / or private information is desensitized by the model desensitization in the desensitization component. In an embodiment of the present application, the desensitization processing method can be to replace the confidential information and / or private information by setting an identifier.

[0062] For example, in an embodiment of the present application, the data desensitizing system includes an identifier pool, which includes different set identifiers, and each set identifier is different. The first query request is "Is the drone model displayed by China at the electronic equipment exhibition M-1D?" First, the data desensitizing system will identify "China", "model", and "M-1D" in the first query request, and then randomly select 3 set identifiers from the identifier pool, namely X1, X2, and X3. Use these three identifiers to replace "China", "model", and "M-1D" in the first query request to obtain a second query request, and the result of the second query request is "Is the drone X2 displayed by X1 at the electronic equipment exhibition X3?"

[0063] After replacing the target with the set identifier, the data desensitization system will also create a mapping relationship between the target information and the set identifier, as shown in Table 1:

[0064] Table 1

[0065] Target Information Setting Identifiers X1 China X2 model X3 M-1D

[0066] After obtaining the mapping relationship shown in Table 1, the data desensitization system will save the mapping relationship to the target database, so that the mapping relationship can be called at any time to translate subsequent query results.

[0067] In an optional embodiment, after obtaining the second query request, the data desensitization system will further determine whether the desensitization processing of the first query request is effective through the discrimination component in the adversarial desensitization module.

[0068] Specifically, the data desensitization system sends the second query request to the determination component, and the determination component performs inference operations based on the information in the second query request to obtain the inference information corresponding to the set identifier in the second query request. It is further determined whether the inference information is consistent with the target information corresponding to the set identifier in the mapping relationship. If they are consistent, it means that the content replaced by the set identifier can be directly inferred through the second query request, so the desensitization processing is invalid, and the desensitization rules used to desensitize the query request need to be updated according to the inference results. If they are inconsistent, it means that the content replaced by the set identifier cannot be inferred through the second query request, so the desensitization rules are valid and there is no need to update the desensitization rules.

[0069] For example, the second query request is "Is the drone X2 displayed by X1 in the electronic equipment exhibition X3?" At this time, the second query request is imported into the discrimination component, and the discrimination component will perform an inference operation on the second query request, that is, infer the content represented by X1, the content represented by X2, and the content represented by X3 in the second query request. If the content represented by X1 is inferred to be "China", the content represented by the set identifier X1 is determined through the above-mentioned mapping relationship. For example, as shown in Table 1, the set identifier X1 is represented as "China", which means that the content inferred by the discrimination component is consistent with the content in the actual mapping relationship. At this time, the desensitization rules in the desensitization component are invalid, so the desensitization rules need to be updated. Thereby, the desensitization rules can more accurately desensitize confidential information or privacy information in the query information.

[0070] In an optional embodiment, after the target information is replaced with the set identifier, the anonymization log component in the adversarial desensitization module will establish a mapping relationship between the target information and the set identifier, and record the mapping relationship, so that after the LLM generates the query results, the query results can be accurately de-anonymized to obtain complete information that can be used by the end user.

[0071] After completing the above-mentioned desensitizing processing of the first query request to obtain the second query request, it is necessary to determine the API key, which is the key for API disconnection verification. That is to say, the second query request will be sent to the server providing LLM service through the API port, and the API key verification is required at this time.

[0072] In the embodiment of the present application, the proxy scheduling module includes an API key pool, a log and monitoring component, and an anonymous communication component. Figure 4 The diagram shows the structure of the proxy scheduling module. The key pool includes an API key set, and the API key set includes various API keys. These API keys correspond to LLM models, that is, each LLM model corresponds to a different API key. Therefore, after the second query request is obtained. The log and monitoring component in the proxy scheduling module will determine the available API keys in the API key set, and detect whether there is an API key corresponding to the LLM model selected by the user. If so, the API key corresponding to the LLM model will be retrieved; if not, any API key will be retrieved from the available API keys. In this way, it can be ensured that the second query request passes the verification of the corresponding API port.

[0073] In addition, it should be noted that the API key pool will regularly add, delete or rotate the API keys in the API key collection. For example, if a new server that can provide LLM services is added, the LLM corresponding to the server and the API key will be saved in the API key pool. Of course, if a server stops providing LLM services, the API key corresponding to the LLM will also be deleted from the key pool. If the LLM service provided by a server is changed, the API key and the corresponding LLM in the API key pool need to be rotated accordingly, so that the validity of each API key can be guaranteed.

[0074] Based on the above process, the proxy scheduling module will send the second query request and the determined API key to the server providing LLM model service through the anonymous communication component, and then the server will verify the API key. After passing the API key verification, the server will output the query result corresponding to the second query request.

[0075] After receiving the query result from the server, the mapping relationship in the target database is retrieved again, and then the setting identifier in the second query request is converted into the corresponding target information, thereby forming the final query result, which is finally sent to the receiving end.

[0076] Through the above method, the target information can be identified in the query request, and then the target information is replaced with the set identifier, and the query request containing the set identifier is sent to the server providing LLM, thus ensuring the security of the data. In addition, in this method, a discriminant component can be used to confirm whether the desensitization process is effective, thereby further improving the security of confidential information.

[0077] The technical solution of this application is further explained below through specific application scenarios.

[0078] Reference Figure 5 The flowchart shown is a workflow of data query. First, the user sends a query request and selects the LLM model. The adversarial desensitization module first desensitizes the query request. The desensitization process has been described in detail in the above embodiment and will not be repeated here.

[0079] After completing the desensitization process, the adversarial desensitization module sends the desensitized query request and the selected LLM model to the proxy scheduling module, and the proxy scheduling module determines the corresponding API key based on the selected LLM model.

[0080] The proxy scheduling module sends the desensitized query request and API key to the server providing the LLM model service. The server will verify the API key. After the verification is passed, the server will output the corresponding query result according to the query request.

[0081] The server returns the query result to the proxy scheduling module, which returns the query result to the adversarial desensitization module, which restores the query result, that is, converts the set identifier in the query result into target information. Finally, the adversarial desensitization module sends the restored query result to the user end.

[0082] Based on the same inventive concept, an information desensitization system is also provided in the embodiment of the present application, referring to Figure 6 As shown, the system includes:

[0083] A receiving module 601 is configured to receive a first query request from a user, wherein the first query request at least includes a request content and a target large language model LLM for analyzing the request content;

[0084] Processing module 602 is used to identify each target information in the first query request according to the identification rule, convert each target information into a set identifier, and obtain a second query request containing the set identifier; retrieve the application programming interface API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service, wherein the target information is confidential information and / or privacy information.

[0085] In an optional embodiment, the processing module 602 is further used to create a mapping relationship between the target information and the corresponding setting identifier; and save the corresponding relationship to the target database.

[0086] In an optional embodiment, the processing module 602 is also used to infer the set identifier in the second query request through an inference model to obtain inference information corresponding to the set identifier; determine whether the inference information is consistent with the target information corresponding to the set identifier in the mapping relationship; if consistent, update the desensitization rules used to desensitize the query request according to the inference result; if inconsistent, maintain the desensitization rules.

[0087] Based on the same inventive concept, an electronic device is also provided in the embodiment of the present application, and the electronic device can realize the functions of the aforementioned information desensitization system, referring to Figure 7 , the electronic device comprises:

[0088] At least one processor 701, and a memory 702 connected to the at least one processor 701. The specific connection medium between the processor 701 and the memory 702 is not limited in the embodiment of the present application. Figure 7 In the example, the processor 701 and the memory 702 are connected via a bus 700. Figure 7 The connections between other components are shown in bold lines, and are not intended to be limiting. The bus 700 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, Figure 7 Only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 701 can also be called a controller, and there is no limitation on the name.

[0089] In the embodiment of the present application, the memory 702 stores instructions that can be executed by at least one processor 701. The at least one processor 701 can execute an information desensitization method discussed above by executing the instructions stored in the memory 702. The processor 701 can implement Figure 6 The functions of each module in the system are shown.

[0090] Among them, the processor 701 is the control center of the device, and can use various interfaces and lines to connect the various parts of the entire control device. By running or executing instructions stored in the memory 702 and calling data stored in the memory 702, the various functions of the device and process data, the device can be monitored as a whole.

[0091] In one possible design, the processor 701 may include one or more processing units, and the processor 701 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the modem processor may not be integrated into the processor 701. In some embodiments, the processor 701 and the memory 702 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on separate chips.

[0092] Processor 701 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of a method for desensitizing information disclosed in conjunction with the embodiments of the present application can be directly embodied as a hardware processor to execute, or a combination of hardware and software modules in the processor to execute.

[0093] The memory 702 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 702 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (Random Access Memory, RAM), a static random access memory (Static Random Access Memory, SRAM), a programmable read-only memory (Programmable Read Only Memory, PROM), a read-only memory (Read Only Memory, ROM), an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a magnetic memory, a disk, an optical disk, etc. The memory 702 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 702 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0094] By programming the processor 701, the code corresponding to the information desensitization method described in the above embodiment can be fixed into the chip, so that the chip can execute the code when running. Figure 1 The steps of an information desensitization method in the embodiment shown are as follows: How to design and program the processor 701 is a technique known to those skilled in the art and will not be described in detail here.

[0095] Based on the same inventive concept, an embodiment of the present application also provides a storage medium, which stores computer instructions. When the computer instructions are executed on a computer, the computer executes an information desensitization method discussed above.

[0096] In some possible implementations, various aspects of an information desensitization method provided by the present application may also be implemented in the form of a program product, which includes a program code. When the program product is run on an apparatus, the program code is used to enable the control device to execute the steps of an information desensitization method according to various exemplary implementations of the present application described above in this specification.

[0097] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0098] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0099] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0100] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0101] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. An information desensitization method, characterized in that: The method comprises: Receive a first query request from a user, wherein the first query request at least includes a request content and a target large language model LLM for analyzing the request content; According to the identification rule, each target information is identified in the first query request, wherein the target information is confidential information and / or private information; Convert each of the target information into a setting identifier to obtain a second query request including the setting identifier; Retrieve the application programming interface API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service.

2. The method according to claim 1, characterized in that After converting each of the target information into a setting identifier, the method further includes: Create a mapping relationship between target information and corresponding setting identifiers; The corresponding relationship is saved in the target database.

3. The method according to claim 2, characterized in that After obtaining the second query request including the set identifier, the method further includes: Inferring the set identifier in the second query request through the inference model to obtain inference information corresponding to the set identifier; Determining whether the inference information is consistent with the target information corresponding to the setting identifier in the mapping relationship; If they are consistent, the desensitization rules for desensitizing the query request are updated according to the inference result; If there is no consistency, the desensitization rule is maintained.

4. The method according to claim 2, characterized in that Create a correspondence between target information and corresponding setting identifiers, including: Retrieving an identifier database containing various types of setting identifiers; A corresponding setting identifier is allocated to each target information, and a mapping relationship between the target information and the setting identifier is established.

5. The method according to claim 1, characterized in that The step of retrieving the application programming interface API key corresponding to the LLM model includes: Determine the available API keys in the API key collection; Detecting whether there is an API key corresponding to the LLM model; If it exists, retrieve the API key corresponding to the LLM model; If it does not exist, retrieve any API key from the available API keys.

6. An information desensitization system, characterized in that: The system comprises: A receiving module, configured to receive a first query request from a user, wherein the first query request at least includes a request content and a target large language model LLM for analyzing the request content; A processing module is used to identify each target information in the first query request according to the identification rule, convert each target information into a set identifier, and obtain a second query request containing the set identifier; retrieve the application programming interface API key corresponding to the LLM model, and send the second query request and the API key to the server providing the LLM model service, wherein the target information is confidential information and / or privacy information.

7. The system according to claim 6, characterized in that The processing module is further used to create a mapping relationship between the target information and the corresponding setting identifier; and save the corresponding relationship to the target database.

8. The system according to claim 7, characterized in that The processing module is further configured to infer the setting identifier in the second query request through an inference model to obtain inference information corresponding to the setting identifier; and determine whether the inference information is consistent with the target information corresponding to the setting identifier in the mapping relationship; If they are consistent, the desensitization rules for desensitizing the query request are updated according to the inference result; if they are inconsistent, the desensitization rules are maintained.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to implement the method steps of any one of claims 1 to 5 when executing the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1 to 5 are implemented.

Images