A method, equipment and product for detecting similar risks

By detecting the similarity between the source code and installation package of software applications across multiple dimensions and combining cross-validation, the problem of detecting similarity risks when software applications are exported overseas is solved, improving the accuracy and compliance of the review and avoiding the triggering of similarity clauses in app stores.

CN123114540APending Publication Date: 2026-11-17QILIN HESHENG NETWORK TECH INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611257179.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-18
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the current technology, there is a lack of effective similarity risk detection methods before software applications are launched overseas. This makes it easy for application matrix packages to trigger similar clauses during the review of overseas app stores, resulting in the rejection of individual applications or the banning of developer accounts.

Method used

By obtaining the application's source code and installation package, similarity detection is performed across multiple dimensions, including resources, hard-coded data, AST structure, functional semantics, SDK architecture, and UI visuals. Combined with cross-validation of the source code and installation package, a similarity risk correction coefficient and risk score are determined to provide an accurate similarity risk assessment.

Benefits of technology

It improves the accuracy of app similarity risk detection, helps developers identify and modify similar risks, avoids triggering similar terms in overseas app stores, and ensures compliance review approval.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123114540A_ABST
    Figure CN123114540A_ABST
Patent Text Reader

Abstract

This application discloses an application similarity risk detection method, device, and product, relating to the field of application compliance auditing technology. The method includes: obtaining the source code and installation package of a first application and the source code and installation package of a second application; detecting the similarity between the source code of the first application and the source code of the second application in multiple dimensions to obtain source code similarity detection results for each dimension; detecting the similarity between the installation package of the first application and the installation package of the second application in each dimension to obtain installation package similarity detection results for each dimension; performing cross-validation on the source code similarity detection results and installation package similarity detection results for each dimension to determine similarity risk correction coefficients and similar risks for each dimension based on the cross-validation results; and determining the similarity risk between the first application and the second application based on the similarity risk correction coefficients and similar risks for each dimension.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of application compliance auditing technology, and in particular to an application-similar risk detection method, equipment, and product. Background Technology

[0002] In recent years, exporting software applications has become a common business model. Before exporting, software applications need to undergo compliance review. Compliance review with overseas app stores is a crucial step in the overall compliance process. In related technologies, some developers or companies often use "application matrix packages" to run multiple applications. Overseas app stores have established clear review rules for this, including clauses prohibiting the submission of applications that lack substantial differences. Once a violation is determined, at best a single application will be rejected, and at worst, it will trigger a related account ban at the developer level, severely impacting the matrix-based operation. Therefore, a technical solution is needed to detect the similarity risk of applications within an application matrix package, providing data support for the compliance review of application matrix packages with overseas app store policies and preventing the triggering of application similarity clauses after the application matrix package is submitted to overseas app stores. Summary of the Invention

[0003] This application provides a method, device, and product for detecting application similarity risks, which can accurately detect application similarity risks.

[0004] In a first aspect, embodiments of this application provide a method for detecting similar risks, the method comprising: Obtain the source code and installation package of the first application and the source code and installation package of the second application; The source code of the first application and the source code of the second application are detected in multiple dimensions to obtain the source code similarity detection results in each dimension. The similarity between the installation package of the first application and the installation package of the second application is detected in each dimension to obtain the installation package similarity detection results in each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results for each dimension, so as to determine the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results for each dimension; Based on the similarity risk correction coefficients and similar risks of each dimension, the similarity risks between the first application and the second application are determined.

[0005] In one embodiment, before obtaining the source code and installation package of the first application and the source code and installation package of the second application, the method further includes: Retrieve all application pairs in the target application matrix package; Each of the aforementioned application pairs is sequentially identified as the target application pair; One application in the target application pair is identified as the first application, and the other application is identified as the second application.

[0006] In one embodiment, the cross-validation of the source code similarity detection results and installation package similarity detection results for each dimension includes: For each of the aforementioned dimensions, the following processing is performed to obtain the cross-validation results for that dimension: Based on the source code similarity detection results of the aforementioned dimension, the source code similarity risk level of the aforementioned dimension is determined in the first level and the second level. Based on the installation package similarity detection results of the aforementioned dimension, the installation package similarity risk level of the aforementioned dimension is determined in the first level and the second level; If the source code similarity risk level is consistent with the installation package similarity risk level, the cross-validation result of the dimension is determined to be consistent with the detection result. If the source code similarity risk level is inconsistent with the installation package similarity risk level, the cross-validation result of the dimension is determined to be inconsistent.

[0007] In one embodiment, determining the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results includes: For each of the dimensions, the following processing is performed to determine the similarity risk correction coefficient and similarity risk for that dimension: If the cross-validation results of the dimension are consistent, the preset similarity risk correction coefficient of the dimension is determined as the similarity risk correction coefficient of the dimension, and the similarity risk of the dimension is determined based on the source code similarity and installation package similarity of the dimension. If the cross-validation results of the dimension are inconsistent, the preset similarity risk correction coefficient of the dimension is adjusted based on the type of inconsistency to obtain the similarity risk correction coefficient of the dimension, and the similarity risk of the dimension is determined based on the installation package similarity of the dimension.

[0008] In one embodiment, the second level is greater than the first level, and a higher level indicates a higher similarity risk; adjusting the preset similarity risk correction coefficient of the dimension based on the type of inconsistency to obtain the similarity risk correction coefficient of the dimension includes: When the inconsistency type is that the source code similarity risk level is level 1 but the installation package similarity risk level is level 2, the preset similarity risk correction coefficient of the dimension is increased to obtain the similarity risk correction coefficient of the dimension. When the inconsistency type is such that the source code similarity risk level is level two but the installation package similarity risk level is level one, the preset similarity risk correction coefficient of the dimension is lowered to obtain the similarity risk correction coefficient of the dimension.

[0009] In one embodiment, the similarity risk includes: a similarity risk score; determining the similarity risk between the first application and the second application based on the similarity risk correction coefficients and similarity risks of each dimension includes: Based on the similarity risk correction coefficients of each dimension, the similarity risk weights of each dimension are corrected to obtain the corrected similarity risk weights of each dimension. Based on the modified similarity risk weights of each dimension, the similarity risk scores of each dimension are weighted and summed to obtain the similarity risk scores of the first application and the second application.

[0010] In one embodiment, the method further includes: If the similarity risk scores of the first application and the second application exceed the application similarity risk score threshold, obtain similar data between the first application and the second application. The similar data is input into a large language model to obtain modification suggestions for the first application and the second application; Output the proposed modifications.

[0011] In one embodiment, the multiple dimensions include: resource dimension, hard-coded dimension, abstract syntax tree (AST) structure dimension, functional semantic dimension, software development kit (SDK) architecture dimension, and user interface (UI) visual dimension. The process of detecting the similarity between the source code of the first application and the source code of the second application across multiple dimensions to obtain source code similarity detection results for each dimension includes: The source code of the first application and the source code of the second application are parsed respectively to obtain the data of each dimension of the source code of the first application and the data of each dimension of the source code of the second application. For each dimension, the data of the dimension of the first application source code and the data of the dimension of the second application source code are input into the similarity detection capability unit Skill corresponding to the dimension to obtain the source code similarity detection result of the dimension. The step of detecting the similarity between the installation packages of the first application and the second application in each dimension to obtain the installation package similarity detection results in each dimension includes: The installation packages of the first application and the second application are parsed respectively to obtain the data of each dimension of the first application installation package and the data of each dimension of the second application installation package; For each dimension, the data of the first application installation package in that dimension and the data of the second application installation package in that dimension are input into the similarity detection capability unit Skill corresponding to that dimension to obtain the installation package similarity detection result for that dimension.

[0012] Secondly, embodiments of this application provide an electronic device, including: A memory on which computer programs are stored; A processor for executing the computer program to implement the above-described application-similar risk detection method.

[0013] Thirdly, embodiments of this application provide a computer-readable program product, including a computer program that can be executed by a processor to implement the above-described application-similar-risk detection method.

[0014] The technical solution provided in this application embodiment obtains the source code and installation package of a first application and the source code and installation package of a second application, detects the similarity between the source code of the first application and the source code of the second application in multiple dimensions to obtain source code similarity detection results in each dimension; detects the similarity between the installation package of the first application and the installation package of the second application in each dimension to obtain installation package similarity detection results in each dimension, performs cross-validation on the source code similarity detection results and installation package similarity detection results in each dimension, determines the similarity risk correction coefficient and similarity risk in each dimension based on the cross-validation results in each dimension, and determines the similarity risk between the first application and the second application based on the similarity risk correction coefficient and similarity risk in each dimension. This allows for a comprehensive assessment of application similarity risk through dual-layer detection and cross-validation of source code and installation package in multiple dimensions, thereby improving the accuracy of application similarity risk detection. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an application of a similarity risk detection method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a similar risk detection device provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0017] The term "first" in the specification and claims of this application The terms "second," etc., are used to distinguish similar objects, not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and "first"... The "second" distinction usually refers to a class of objects, without limiting the number of objects. For example, the first object can be one or more. In addition, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0018] This application provides a method for detecting application similarity risks. This method accurately detects application similarity risks within an application matrix package, providing data support for developers' internal compliance review of application matrix packages for overseas app store policies. This prevents the application matrix package from triggering application similarity clauses in overseas app stores after submission. This method can be applied to the overseas compliance review platform within the development entity and executed by that platform.

[0019] Figure 1 This is a schematic diagram of a similarity risk detection process provided in an embodiment of this application. For example... Figure 1 As shown, the process includes: S102, Obtain the source code and installation package of the first application and the source code and installation package of the second application.

[0020] The first application and the second application can be any two different applications from the target application matrix package. The target application matrix package is the application matrix package to be subjected to application similarity risk detection, which can include multiple applications. The applications are software applications, such as mobile apps.

[0021] In practice, developers can submit a target application matrix package and the source code repository addresses of each application within the target application matrix package to the overseas compliance review platform. The target application matrix package includes the installation packages of each application. The overseas compliance review platform can obtain the installation packages of the first and second applications from the target application matrix package, and obtain the source code of the first application based on its source code repository address, and the source code of the second application based on its source code repository address.

[0022] Similarity between any two applications in the target application matrix will trigger the application similarity risk clause of the app store. In one embodiment, to perform comprehensive similarity risk detection on the target application matrix, the process may further include the following before step S102: Retrieve all application pairs in the target application matrix package; sequentially identify each application pair as the target application pair; identify one application in the target application pair as the first application and the other application as the second application.

[0023] In practice, the target application matrix package can be parsed to obtain the application list of the target application matrix. By exhausting all pairwise combinations of applications in the application list, all application pairs in the target application matrix package can be obtained.

[0024] For example, assuming the parsed target application matrix package contains application 1, application 2, and application 3, we can exhaustively combine all pairs of applications in the matrix package to obtain application pairs in the matrix package: (application 1, application 3), (application 2, application 3), and (application 1, application 3). Each of these application pairs is then taken as the target application pair, with one application in the target application pair designated as the first application and the other as the second application. Thus, application similarity risk detection is performed on each application pair.

[0025] S104, detect the similarity between the source code of the first application and the source code of the second application in each of multiple dimensions to obtain the source code similarity detection results in each dimension.

[0026] Among these dimensions, multiple dimensions can be preset dimensions. In one embodiment, multiple dimensions may include the following dimensions: resource dimension, hard-coded dimension, abstract syntax tree (AST) structural dimension, functional semantic dimension, software development kit (SDK) architectural dimension, and user interface (UI) visual dimension.

[0027] Similarity in the resource dimension refers to the similarity at the level of non-code resource files. Non-code resource files are static resource files, including but not limited to: images, audio, configuration files, etc. Similarity in the hard-coded dimension refers to the similarity at the level of hard-coded content. Hard-coded content consists of directly defined constant values, including but not limited to: keys, interface addresses, configuration switches, and event tracking domains, etc. Similarity in the AST structure dimension refers to the similarity at the level of code syntax. Similarity in the functional semantic dimension refers to the similarity at the level of code functionality. Similarity in the SDK architecture dimension refers to the similarity at the level of project architecture and dependent frameworks. Similarity in the UI visual dimension refers to the similarity at the level of visual presentation elements. Visual presentation elements include but are not limited to: key interface screenshots, control arrangement, page structure, etc.

[0028] In one embodiment, the source code of the first application and the source code of the second application can be parsed to obtain data of each dimension of the source code of the first application and the source code of the second application respectively; for each dimension, the data of that dimension of the source code of the first application and the source code of the second application are input into the similarity detection capability unit Skill corresponding to that dimension to obtain the source code similarity detection result for that dimension.

[0029] For example, the source code of the first application can be parsed to obtain its resource files, hard-coded content, AST structure, code functional semantic vectors, SDK architecture, and UI visual elements. Similarly, the source code of the second application can be parsed to obtain the same resource files, hard-coded content, AST structure, code functional semantic vectors, SDK architecture, and UI visual elements. Then, the resource files of both applications can be input into a resource similarity detection skill for similarity testing. The results obtained represent the source code similarity of the first and second applications in terms of resource dimensions. Similarly, the hard-coded content of both applications can be input into a hard-coded similarity detection skill for similarity testing. The results obtained represent the source code similarity of the first and second applications in terms of hard-coded dimensions. The AST structures of both applications can be input into an AST structure similarity detection skill for similarity testing. The results obtained represent the source code similarity of the first and second applications in terms of AST structure dimensions. The semantic vectors of the source code functions and the semantic vectors of the second application's source code are input into the semantic similarity detection skill to perform similarity detection, and the detection result is the source code similarity detection result of the semantic dimension of the first application and the second application. The SDK architecture of the first application's source code and the SDK architecture of the second application's source code are input into the SDK architecture similarity detection skill to perform similarity detection, and the detection result is the source code similarity detection result of the SDK architecture dimension of the first application and the second application. The UI visual elements of the first application's source code and the UI visual elements of the second application's source code are input into the UI visual similarity detection skill to perform similarity detection, and the detection result is the source code similarity detection result of the UI visual dimension of the first application and the second application.

[0030] Among them, the similarity detection capability unit Skill for each dimension is a professional capability module that is specifically responsible for similarity detection in that dimension and can be run independently / plugged in. It is a complete functional entity that encapsulates the algorithm model, detection logic, and judgment rules for similarity detection in that dimension. In implementation, it can be implemented based on the model.

[0031] S106, detect the similarity between the installation package of the first application and the installation package of the second application in each dimension to obtain the installation package similarity detection results in each dimension.

[0032] In one embodiment, the installation packages of the first application and the second application can be parsed separately to obtain data for each dimension of the first application installation package and data for each dimension of the second application installation package; for each dimension, the data for that dimension of the first application installation package and the data for that dimension of the second application installation package are input into the similarity detection capability unit Skill corresponding to that dimension to obtain the installation package similarity detection result for that dimension.

[0033] For example, the installation package of the first application can be parsed to obtain its resource files, hard-coded content, AST structure, code functional semantic vectors, SDK architecture, and UI visual elements. Similarly, the installation package of the second application can be parsed to obtain the same resource files, hard-coded content, AST structure, code functional semantic vectors, SDK architecture, and UI visual elements. Then, the resource files of both the first and second application installation packages can be input into a resource similarity detection skill for similarity testing. The detection result represents the resource-level similarity of the installation packages for both applications. Similarly, the hard-coded content of both applications can be input into a hard-coded similarity detection skill for similarity testing. This result represents the hard-coded similarity of the installation packages for similarity testing. Likewise, the AST structures of both applications can be input into an AST structure similarity detection skill for similarity testing. This result represents the AST structure-level similarity of the installation packages for similarity testing. The semantic vectors of the code functions of the first application's installation package and the second application's installation package are input into the semantic similarity detection skill to perform similarity detection, and the detection result is the semantic similarity detection result of the first application and the second application's installation packages. The SDK architecture of the first application's installation package and the SDK architecture of the second application's installation package are input into the SDK architecture similarity detection skill to perform similarity detection, and the detection result is the SDK architecture similarity detection result of the first application and the second application's installation packages. The UI visual elements of the first application's installation package and the UI visual elements of the second application's installation package are input into the UI visual similarity detection skill to perform similarity detection, and the detection result is the UI visual similarity detection result of the first application and the second application's installation packages.

[0034] It is understandable that steps S104 and S106 are independent of each other, and they are detections of two channels that can be executed in parallel.

[0035] S108, perform cross-validation on the source code similarity detection results and installation package similarity detection results of each dimension, so as to determine the similarity risk correction coefficient and similarity risk of each dimension based on the cross-validation results of each dimension.

[0036] The cross-validation results for each dimension can include whether the detection results for that dimension are consistent or inconsistent.

[0037] The cross-validation for each dimension is independent. In one embodiment, for each dimension, the following processing can be performed to obtain the cross-validation result for that dimension: Based on the source code similarity detection results for this dimension, the source code similarity risk level for this dimension is determined in the first and second levels; based on the installation package similarity detection results for this dimension, the installation package similarity risk level for this dimension is determined in the first and second levels; if the source code similarity risk level and the installation package similarity risk level are consistent, the cross-validation result for this dimension is determined to be consistent; if the source code similarity risk level and the installation package similarity risk level are inconsistent, the cross-validation result for this dimension is determined to be inconsistent.

[0038] In this system, the second level can be higher than the first level; a higher level indicates a greater risk of similarity. In other words, the first level represents a low risk of similarity, and the second level represents a high risk of similarity. If both the source code similarity risk level and the installation package similarity risk level for this dimension are at the first level or both are at the second level, the cross-validation results for this dimension are considered consistent. If the source code similarity risk level for this dimension is first but the installation package similarity risk level is second, or vice versa, the cross-validation results for this dimension are considered inconsistent.

[0039] The risk levels of source code similarity and installation package similarity in this dimension can be determined using appropriate methods.

[0040] In one embodiment, the source code similarity detection result for this dimension includes: the source code similarity score for this dimension, and the installation package similarity detection result for this dimension includes: the installation package similarity score for this dimension. The source code similarity score for this dimension can be compared with a scoring threshold for this dimension. If the source code similarity score for this dimension is greater than the scoring threshold, the source code similarity risk level for this dimension is determined to be Level 2; otherwise, the source code similarity risk level for this dimension is determined to be Level 1. Similarly, the installation package similarity score for this dimension can be compared with a scoring threshold for this dimension. If the installation package similarity score for this dimension is greater than the scoring threshold, the installation package similarity risk level for this dimension is determined to be Level 2; otherwise, the installation package similarity risk level for this dimension is determined to be Level 1. The scoring thresholds for different dimensions can be different.

[0041] In one embodiment, the source code similarity detection result for this dimension may include: a source code similarity score and a source code similarity risk level for this dimension, wherein the source code similarity risk level is either level one or level two. The installation package similarity detection result for this dimension may include: an installation package similarity score and an installation package similarity risk level for this dimension, wherein the installation package similarity risk level is either level one or level two. The source code similarity risk level for this dimension can be directly obtained from the source code similarity detection result for this dimension, and the installation package similarity risk level for this dimension can be obtained directly from the installation package similarity detection result for this dimension.

[0042] The similarity risk correction coefficients for each dimension are used to correct the similarity risk weights for each dimension. These corrected similarity risk weights are then used to weight the similarity risks across dimensions to determine the application similarity risk. Different validation results may correspond to different similarity risk correction coefficients and similarity risk determination strategies.

[0043] In one embodiment, for each of the dimensions, the following processing can be performed to determine the similarity risk correction factor and similarity risk for that dimension: If the cross-validation results for this dimension are consistent, the preset similarity risk correction coefficient for this dimension is determined as the similarity risk correction coefficient for this dimension, and the similarity risk for this dimension is determined based on the source code similarity and installation package similarity for this dimension. If the cross-validation results for this dimension are inconsistent, the preset similarity risk correction coefficient for this dimension is adjusted based on the type of inconsistency to obtain the similarity risk correction coefficient for this dimension, and the similarity risk for this dimension is determined based on the installation package similarity for this dimension.

[0044] The similarity risk in this dimension indicates the risk that similarity in this dimension may lead to app similarity, thereby triggering the app similarity terms in app stores. Similarity risk is positively correlated with similarity; the higher the similarity, the higher the similarity risk, and vice versa.

[0045] In one implementation, similarity is specifically a similarity score. The source code similarity detection result for this dimension includes the source code similarity score for this dimension, the installation package similarity detection result for this dimension includes the installation package similarity score for this dimension, and similarity risk is specifically a similarity risk score.

[0046] If the cross-validation results for this dimension are consistent, a weighted sum of the source code similarity score and the installation package similarity score for this dimension can be performed based on the source code similarity weight and the installation package similarity weight to obtain the similarity risk score for this dimension. The source code similarity weight represents the impact of the source code similarity score on the similarity risk score for this dimension, and the installation package similarity weight represents the impact on the similarity risk score for this dimension. The source code similarity weight and the installation package weight are preset weights; a larger weight indicates a greater impact, and a smaller weight indicates a smaller impact.

[0047] If the cross-validation results for this dimension are inconsistent, the installation package similarity score for this dimension can be determined as the similarity risk score for that dimension. After an application is submitted to an app store, the app store typically cannot obtain the original source code. Its similarity determination is usually based on the installation package and its decompiled products. When the source code similarity detection results are inconsistent with the installation package similarity detection results, determining the similarity risk score based on the installation package similarity score can ensure that the developer's internal application similarity risk detection standards are consistent with the app store's actual judgment perspective, avoiding missed or false positives.

[0048] The similarity risk correction coefficient for this dimension is used to correct the similarity risk weights for this dimension.

[0049] In one embodiment, the similarity risk correction coefficient for this dimension is used to linearly amplify the similarity risk weights for that dimension. The larger the similarity risk correction coefficient, the greater the amplification effect on the similarity risk weights; the smaller the correction coefficient, the smaller the amplification effect on the similarity risk weights.

[0050] When the inconsistent type has a source code similarity risk level of Level 1 but an installation package similarity risk level of Level 2, the preset similarity risk correction coefficient for that dimension can be increased to obtain the corrected similarity risk coefficient for that dimension; when the inconsistent type has a source code similarity risk level of Level 2 but an installation package similarity risk level of Level 1, the preset similarity risk correction coefficient for that dimension can be decreased to obtain the corrected similarity risk coefficient for that dimension.

[0051] The system is divided into two levels: Level 1 represents low similarity risk, and Level 2 represents high similarity risk. The system adjusts the preset similarity risk correction coefficient for a given dimension by increasing it when the source code similarity risk level is low but the installation package similarity risk level is high, and by decreasing it when the source code similarity risk level is high but the installation package similarity risk level is low. This approach uses the installation package perspective as the anchor point for correction, ensuring consistency between the application similarity risk detection standards used by developers for internal compliance reviews and the actual judgment perspective of app stores. This allows developers to accurately predict the actual actions of app stores based on their internal application similarity risk detection results, enabling them to make proactive rectifications.

[0052] In implementation, the magnitude of the upward and downward adjustments can be determined based on the installation package similarity score for that dimension. When the inconsistency type is a source code similarity risk level of Level 1 but an installation package similarity risk level of Level 2, a higher installation package similarity score allows for a larger upward adjustment. Conversely, when the inconsistency type is a source code similarity risk level of Level 2 but an installation package similarity risk level of Level 1, a lower installation package similarity score allows for a larger downward adjustment.

[0053] For example, the adjustment range can be determined based on the distance between the installation package similarity score and the score threshold. In cases where the source code similarity risk level is Level 1 but the installation package similarity risk level is Level 2, the similarity risk correction coefficient for that dimension can be determined using the following formula.

[0054]

[0055] in, For the first Similarity risk correction coefficients in each dimension For the first The installation package similarity score is calculated across several dimensions. For the first Scoring thresholds for each dimension, This is the preset adjustment step size.

[0056] In cases where the source code similarity risk level is Level 2 but the installation package similarity risk level is Level 1, the similarity risk correction coefficient for that dimension can be determined based on the following formula.

[0057]

[0058] in, For the first Similarity risk correction coefficients in each dimension For the first The installation package similarity score is calculated across several dimensions. For the first Scoring thresholds for each dimension, This is the preset adjustment step size.

[0059] S110, based on the similarity risk correction coefficients and similar risks of each dimension, determine the similarity risks between the first application and the second application.

[0060] In one embodiment, similar risk specifically refers to a similar risk score. The similar risk weights for each dimension can be adjusted based on the similar risk adjustment coefficients for each dimension to obtain adjusted similar risk weights for each dimension. Based on these adjusted similar risk weights, the similar risk scores for each dimension are then weighted and summed to obtain the similar risk score between the first application and the second application. The formula is illustrated below:

[0061] in, For the similar risks of the first and second applications, D is a set of multiple dimensions. For dimensional indexing, For the first Similarity risk weights in each dimension For the first Similar risks in each dimension For the first Similarity risk correction coefficients for each dimension.

[0062] The application similarity risk detection method provided in this application embodiment obtains the source code and installation package of a first application and the source code and installation package of a second application. It detects the similarity between the source code of the first application and the source code of the second application across multiple dimensions to obtain source code similarity detection results for each dimension. It also detects the similarity between the installation packages of the first application and the second application across each dimension to obtain installation package similarity detection results for each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results to determine the similarity risk correction coefficient and similarity risk for each dimension. Based on the similarity risk correction coefficient and similarity risk for each dimension, the similarity risk between the first application and the second application is determined. This method can comprehensively evaluate application similarity risk through dual-layer detection and cross-validation of source code and installation packages across multiple dimensions, thereby improving the accuracy of application similarity risk detection. Furthermore, the first and second applications can be any two applications in the target application matrix package. Through the above method, accurate data can be provided for the compliance review of the target application matrix package with overseas app store policies. This makes it easier for developers to modify similar high-risk applications in the target application matrix package and avoid triggering the application similarity clauses of overseas app stores after the target application matrix package is submitted to overseas app stores.

[0063] In one embodiment, to facilitate data support for the overseas app store policy compliance review of the target application matrix package, a similarity risk review report can be generated based on the source code similarity detection results, installation package similarity detection results, cross-validation results of the various dimensions, similarity risk correction coefficients and similar risks of the various dimensions, and the similarity risks of the first application and the second application. The review report can then be output for relevant personnel to review.

[0064] In one embodiment, to facilitate developers in modifying applications with high similarity risk, if the similarity risk scores of the first application and the second application exceed the application similarity risk score threshold, similarity data of the first application and the second application can be obtained; the similarity data can be input into a large language model to obtain modification suggestions for the first application and the second application; and the modification suggestions can be output.

[0065] In implementation, the source code similarity detection results for each dimension can include: the source code similarity score for that dimension and similarity data between the source code of the first application and the source code of the second application for that dimension. The installation package similarity detection results for each dimension can include: the installation package similarity score for that dimension and similarity data between the installation package of the first application and the installation package of the second application for that dimension. The similarity data from the source code similarity detection results and installation package similarity detection results for each dimension can be obtained and input into the large language model. The large language model is then instructed to output modification suggestions for the first and second applications based on this similarity data, while maintaining the original business logic. The modification suggestions input from the large language model can be output to developers so that they can modify the first and / or second applications. These modification suggestions include, but are not limited to: resource variant replacement schemes, UI layout and visual style differentiation adjustment suggestions, and general business logic module modification suggestions.

[0066] After the modifications are completed, the developers can resubmit the modified first and second applications to the overseas compliance testing platform, which can then perform similar risk detection methods on the applications again.

[0067] The above describes an application similarity risk detection method provided in the embodiments of this application.

[0068] Based on the same technical concept as the above embodiments, this application also provides a similar risk detection device. For example... Figure 2 As shown in the figure, an application similarity risk detection device 200 provided in this application embodiment includes: The acquisition module 210 is used to acquire the source code and installation package of the first application and the source code and installation package of the second application; The source code detection module 220 is used to detect the similarity between the source code of the first application and the source code of the second application in multiple dimensions, so as to obtain the source code similarity detection results in each dimension. The installation package detection module 230 is used to detect the similarity between the installation package of the first application and the installation package of the second application in each dimension, so as to obtain the installation package similarity detection results in each dimension. The cross-validation module 240 is used to cross-validate the source code similarity detection results and installation package similarity detection results of each dimension, so as to determine the similarity risk correction coefficient and similarity risk of each dimension based on the cross-validation results of each dimension; The risk determination module 250 is used to determine the similarity risk between the first application and the second application based on the similarity risk correction coefficient and similar risk of each dimension.

[0069] Obviously, the apparatus in the embodiments of this application can be used as described above. Figure 1 The execution entity of the method shown is capable of achieving Figure 1 The function of the method shown is the same, so it will not be repeated here.

[0070] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Please refer to it. Figure 3 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random access memory. Random-Access Memory, RAM It may also include non-volatile memory, such as at least one disk storage device. Of course, the electronic device may also include other hardware required for its operations.

[0071] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA. Industry Standard Architecture (ISA) bus, PCI (Peripheral Component Interconnect) bus or EISA Extended Industry Standard Architecture (ESA) buses, etc. These buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0072] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0073] The processor reads the corresponding computer program from non-volatile memory into main memory and then runs it, forming an application similarity risk detection device at the logical level. The processor executes the program stored in memory and specifically performs the following operations: Obtain the source code and installation package of the first application and the source code and installation package of the second application; The source code of the first application and the source code of the second application are detected in multiple dimensions to obtain the source code similarity detection results in each dimension. The similarity between the installation package of the first application and the installation package of the second application is detected in each dimension to obtain the installation package similarity detection results in each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results for each dimension, so as to determine the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results for each dimension; Based on the similarity risk correction coefficients and similar risks of each dimension, the similarity risks between the first application and the second application are determined.

[0074] The above is as stated in this application. Figure 1 The methods disclosed in the illustrated embodiments can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0075] It should be understood that the electronic device in the embodiments of this application can achieve Figure 1 The illustrated method embodiments demonstrate the functionality of the method. Since the principles are the same, the embodiments of this application will not be described again here.

[0076] Of course, in addition to software implementation, the electronic device of this application does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0077] This application also provides a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by a processor, can perform... Figure 1 The method of the illustrated embodiment is specifically used to perform the following operations: Obtain the source code and installation package of the first application and the source code and installation package of the second application; The source code of the first application and the source code of the second application are detected in multiple dimensions to obtain the source code similarity detection results in each dimension. The similarity between the installation package of the first application and the installation package of the second application is detected in each dimension to obtain the installation package similarity detection results in each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results for each dimension, so as to determine the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results for each dimension; Based on the similarity risk correction coefficients and similar risks of each dimension, the similarity risks between the first application and the second application are determined.

[0078] It should be understood that the computer-readable storage medium of the embodiments of this application can implement Figure 1 The illustrated method embodiments demonstrate the functionality of the method. Since the principles are the same, the embodiments of this application will not be described again here.

[0079] This application also provides a computer program product, including a computer program, which, when executed by a processor, can achieve... Figure 1 The method of the illustrated embodiment is specifically used to perform the following operations: Obtain the source code and installation package of the first application and the source code and installation package of the second application; The source code of the first application and the source code of the second application are detected in multiple dimensions to obtain the source code similarity detection results in each dimension. The similarity between the installation package of the first application and the installation package of the second application is detected in each dimension to obtain the installation package similarity detection results in each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results for each dimension, so as to determine the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results for each dimension; Based on the similarity risk correction coefficients and similar risks of each dimension, the similarity risks between the first application and the second application are determined.

[0080] It should be understood that the computer program product of the embodiments of this application can implement Figure 1 The illustrated method embodiments demonstrate the functionality of the method. Since the principles are the same, the embodiments of this application will not be described again here.

[0081] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0082] In summary, the above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

[0083] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0084] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0085] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0086] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

Claims

1. A method for detecting similar risks, characterized in that, The method includes: Obtain the source code and installation package of the first application and the source code and installation package of the second application; The source code of the first application and the source code of the second application are detected in multiple dimensions to obtain the source code similarity detection results in each dimension. The similarity between the installation package of the first application and the installation package of the second application is detected in each dimension to obtain the installation package similarity detection results in each dimension. Cross-validation is performed on the source code similarity detection results and installation package similarity detection results for each dimension, so as to determine the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results for each dimension; Based on the similarity risk correction coefficients and similar risks of each dimension, the similarity risks between the first application and the second application are determined.

2. The method according to claim 1, characterized in that, Before obtaining the source code and installation package of the first application and the source code and installation package of the second application, the method further includes: Retrieve all application pairs in the target application matrix package; Each of the aforementioned application pairs is sequentially identified as the target application pair; One application in the target application pair is identified as the first application, and the other application is identified as the second application.

3. The method according to claim 1, characterized in that, The cross-validation of the source code similarity detection results and installation package similarity detection results for each dimension includes: For each of the aforementioned dimensions, the following processing is performed to obtain the cross-validation results for that dimension: Based on the source code similarity detection results of the aforementioned dimension, the source code similarity risk level of the aforementioned dimension is determined in the first level and the second level. Based on the installation package similarity detection results of the aforementioned dimension, the installation package similarity risk level of the aforementioned dimension is determined in the first level and the second level; If the source code similarity risk level is consistent with the installation package similarity risk level, the cross-validation result of the dimension is determined to be consistent with the detection result. If the source code similarity risk level is inconsistent with the installation package similarity risk level, the cross-validation result of the dimension is determined to be inconsistent.

4. The method according to claim 3, characterized in that, The determination of the similarity risk correction coefficient and similarity risk for each dimension based on the cross-validation results of each dimension includes: For each of the dimensions, the following processing is performed to determine the similarity risk correction coefficient and similarity risk for that dimension: If the cross-validation results of the dimension are consistent, the preset similarity risk correction coefficient of the dimension is determined as the similarity risk correction coefficient of the dimension, and the similarity risk of the dimension is determined based on the source code similarity and installation package similarity of the dimension. If the cross-validation results of the dimension are inconsistent, the preset similarity risk correction coefficient of the dimension is adjusted based on the type of inconsistency to obtain the similarity risk correction coefficient of the dimension, and the similarity risk of the dimension is determined based on the installation package similarity of the dimension.

5. The method according to claim 4, characterized in that, The second level is higher than the first level, and the higher the level, the higher the similar risk. The step of adjusting the preset similarity risk correction coefficient of the dimension based on the type of inconsistency to obtain the similarity risk correction coefficient of the dimension includes: When the inconsistency type is that the source code similarity risk level is level 1 but the installation package similarity risk level is level 2, the preset similarity risk correction coefficient of the dimension is increased to obtain the similarity risk correction coefficient of the dimension. When the inconsistency type is such that the source code similarity risk level is level two but the installation package similarity risk level is level one, the preset similarity risk correction coefficient of the dimension is lowered to obtain the similarity risk correction coefficient of the dimension.

6. The method according to claim 1, characterized in that, The similarity risk includes: a similarity risk score; determining the similarity risk between the first application and the second application based on the similarity risk correction coefficients and similarity risks of each dimension includes: Based on the similarity risk correction coefficients of each dimension, the similarity risk weights of each dimension are corrected to obtain the corrected similarity risk weights of each dimension. Based on the modified similarity risk weights of each dimension, the similarity risk scores of each dimension are weighted and summed to obtain the similarity risk scores of the first application and the second application.

7. The method according to claim 6, characterized in that, The method further includes: If the similarity risk scores of the first application and the second application exceed the application similarity risk score threshold, obtain similar data between the first application and the second application. The similar data is input into a large language model to obtain modification suggestions for the first application and the second application; Output the proposed modifications.

8. The method according to claim 1, characterized in that, The multiple dimensions include: resource dimension, hard-coded dimension, abstract syntax tree (AST) structure dimension, functional semantic dimension, software development kit (SDK) architecture dimension, and user interface (UI) visual dimension. The process of detecting the similarity between the source code of the first application and the source code of the second application across multiple dimensions to obtain source code similarity detection results for each dimension includes: The source code of the first application and the source code of the second application are parsed respectively to obtain the data of each dimension of the source code of the first application and the data of each dimension of the source code of the second application. For each dimension, the data of the dimension of the first application source code and the data of the dimension of the second application source code are input into the similarity detection capability unit Skill corresponding to the dimension to obtain the source code similarity detection result of the dimension. The step of detecting the similarity between the installation packages of the first application and the second application in each dimension to obtain the installation package similarity detection results in each dimension includes: The installation packages of the first application and the second application are parsed respectively to obtain the data of each dimension of the first application installation package and the data of each dimension of the second application installation package; For each dimension, the data of the first application installation package in that dimension and the data of the second application installation package in that dimension are input into the similarity detection capability unit Skill corresponding to that dimension to obtain the installation package similarity detection result for that dimension.

9. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program to implement the application similarity risk detection method according to any one of claims 1 to 8.

10. A computer-readable program product, characterized in that, It includes a computer program that can be executed by a processor to implement the application similarity risk detection method according to any one of claims 1 to 8.