Cross-border data flow management and control system and method based on privacy calculation verification and medium

Abstract

The invention discloses a cross-border data flow management and control system and method based on privacy calculation verification and a medium. The system comprises a legal knowledge graph construction module, a compliance strategy dynamic generation module, a privacy calculation sandbox module, a data flow topology analysis module and a full-link audit tracing module. Natural language terms are converted into programmable compliance rules through a legal knowledge graph, a technical system is driven to automatically adapt to policy changes, the whole process from single-point protection to full-link privacy calculation is achieved, namely, the whole process of data encryption transmission, secret state calculation and compliance verification is covered, global laws and regulations can be analyzed in real time, and an executable technical strategy can be generated; from sovereign confrontation to sovereign collaboration, a sovereign compliance map is constructed through digital twinning, and a data flow path is optimized. According to the method, digital watermark marking is carried out on all nodes of data flow, traceability and responsibility definition of leakage events can be realized, and a cross-border data flow space-time barrier is broken through.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-border data flow management, and specifically to a cross-border data flow management and control system, method and medium based on privacy computing verification. Background Art

[0002] The main technical means relied upon by mainstream cross-border data flow security management technologies in their actual use have the following problems:

[0003] Traditional encryption transmission technology requires plain text processing within the data receiving domain, which poses a risk of secondary leakage; anonymization technology is susceptible to re-identification attacks and destroys data correlation, affecting analysis accuracy.

[0004] Using localized data storage and physical isolation will sacrifice data flow efficiency and cannot support global business needs.

[0005] Using static compliance review tools and compliance screening systems based on a preset rule base, they rely on manual upload of legal text updates, are unable to respond to policy changes in real time, have insufficient dynamic compliance adaptation capabilities, are prone to legal-technical gaps, and lag in updating the static compliance rule base, which can easily lead to violation risks. At the same time, legal terms are disconnected from technical parameters and rely on manual mapping based on expert experience.

[0006] Anonymization / desensitization technology has difficulty balancing privacy protection and data availability.

[0007] The lack of a cross-border collaborative governance mechanism makes it easy for sovereignty border conflicts to occur, leading to rigid routing and a lack of flexibility and elasticity. At the same time, multi-party collaboration lacks a credible audit mechanism, making it difficult to meet the evidence requirements of multiple jurisdictions. Therefore, cross-border collaboration lacks a unified technical mutual trust foundation and dynamic arbitration mechanism, exacerbating sovereignty border conflicts. Summary of the Invention

[0008] In view of this, embodiments of the present invention provide a cross-border data flow control system, method, and medium based on privacy computing verification.

[0009] The first aspect of the present invention provides a cross-border data flow control system based on privacy computing verification, including a legal knowledge graph construction module, a compliance policy dynamic generation module, a privacy computing sandbox module, a data flow topology analysis module, and a full-link audit tracing module;

[0010] The legal knowledge graph construction module is used to construct a legal knowledge graph and a multi-dimensional weight matrix of laws of multiple countries;

[0011] The compliance strategy dynamic generation module is used to generate a compliance strategy combination based on the legal knowledge graph of multiple countries' laws;

[0012] The privacy computing sandbox module is used to encrypt cross-border data and then pre-verify the encrypted cross-border data using the compliance policy combination in the distributed sandbox;

[0013] The data flow topology analysis module is used to build a real-time topology network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multiple countries' laws into different topology nodes, and use the real-time topology network to control the flow of cross-border data;

[0014] The full-link audit and tracing module is used to monitor the data flow of cross-border data.

[0015] Furthermore, in the legal knowledge graph construction module, a legal knowledge graph and a multi-dimensional weight matrix of laws of multiple countries are constructed, specifically including the following steps:

[0016] Obtain legal texts from multiple countries, generate structured legal metadata based on the legal texts, and map the relationship between each legal metadata to establish a legal knowledge graph;

[0017] The conflict coefficients between the various legal metadata and mapping relationships are determined from the legal knowledge graph, and a multi-dimensional weight matrix of each country's laws is constructed based on the conflict coefficients.

[0018] Furthermore, the legal knowledge graph construction module is further configured to periodically update the legal knowledge graph and the multi-dimensional weight matrix; the periodic updating of the legal knowledge graph and the multi-dimensional weight matrix specifically includes the following steps:

[0019] Regularly obtain the latest versions of legal texts from various countries;

[0020] Detecting semantic changes in newly acquired legal texts from multiple countries; semantic changes include semantic addition, semantic reduction, and semantic shift;

[0021] When semantic changes are found in newly acquired legal texts from multiple countries, the legal knowledge graph is reorganized and conflicts are resolved to obtain a new legal knowledge graph;

[0022] The multi-dimensional weight matrix is ​​updated according to the conflict coefficient of the new legal knowledge graph.

[0023] Furthermore, the conflict resolution specifically includes the following steps:

[0024] Detecting vertical and horizontal conflicts in the reorganized legal knowledge graph; the vertical conflict represents the conflict between the mapping relationships of new and old legal metadata of the same country after the reorganization; the horizontal conflict represents the conflict between the mapping relationships of legal metadata of different countries after the reorganization;

[0025] Resolving the vertical and horizontal conflicts according to a preset resolution strategy; the resolution strategy includes soft resolution and hard resolution; the soft resolution includes adjusting the confidence level of the mapping relationship between legal metadata, and the hard resolution includes establishing or deleting the mapping relationship between legal metadata;

[0026] For conflicts that cannot be resolved by the preset resolution strategy, manual review is introduced.

[0027] Furthermore, in the compliance strategy dynamic generation module, generating a compliance strategy combination based on the legal knowledge graph of multiple countries' laws specifically includes the following steps:

[0028] Query key obligations of different countries’ laws through graph neural networks in the legal knowledge graph of multiple countries’ laws;

[0029] Take the greatest common divisor of key obligations under the laws of different countries as the intersection of key obligations;

[0030] Generate a portfolio of compliance strategies suitable for different countries based on the intersection of key obligations.

[0031] Furthermore, in the privacy computing sandbox module, cross-border data is locally encrypted through the OT-HE combined protocol, and the encrypted gradient data is transmitted between distributed sandboxes;

[0032] The distributed sandbox stores a compliance policy combination, which is used to jointly calculate the encrypted gradient data, and returns the calculation result to the local decryption of the cross-border data as a pre-verification result.

[0033] Furthermore, in the data flow topology analysis module, the real-time topology network is used to control the flow of cross-border data, specifically including the following steps:

[0034] Determine the data path of cross-border data in the topological network, traverse the legal knowledge graph and multi-dimensional weight matrix involved in the data path, and determine the overall compliance risk of the data path;

[0035] When the total compliance risk of the data path exceeds the preset threshold, cross-border data will be blocked or the path will be switched.

[0036] Furthermore, the full-link audit tracing module is used to add digital watermarks to each data processing node of the legal knowledge graph construction module, the compliance policy dynamic generation module, the privacy computing sandbox module and the data flow topology analysis module, and monitor and store the data flow of cross-border data through digital watermarks; and write the stored content into the blockchain.

[0037] Another aspect of the present invention discloses a cross-border data flow control method based on privacy computing verification, comprising the following steps:

[0038] Construct legal knowledge graphs and multi-dimensional weight matrices of laws from multiple countries;

[0039] Generate compliance strategy combinations based on the legal knowledge graph of multiple countries’ laws;

[0040] After encrypting the cross-border data, pre-verify the encrypted cross-border data using the compliance policy combination in the distributed sandbox;

[0041] Construct a real-time topological network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multiple countries' laws into different topological nodes, and use the real-time topological network to control the flow of cross-border data;

[0042] Monitor data flows across borders.

[0043] Yet another aspect of the present invention discloses a computer-readable storage medium, wherein the storage medium stores a program for implementing the system described in the first aspect, and the program can be executed by a processor.

[0044] The present invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the above method.

[0045] The embodiments of the present invention have the following beneficial effects: A cross-border data flow control system, method, and medium based on privacy computing verification of the present invention have at least the following effects:

[0046] (1) This invention uses digital mapping of legal terms, analyzes privacy regulations across languages, constructs a three-dimensional legal knowledge graph (jurisdiction / legal entity / obligation conditions), and establishes a mapping relationship between legal terms and technical parameters. It implements multi-jurisdictional clause conflict detection, forms a machine-readable compliance benchmark rule base, and provides executable legal parameter input for subsequent modules;

[0047] (2) Using data sovereignty boundary modeling, based on IP geo-tags and blockchain transmission logs, real-time tracking of the physical nodes and jurisdictions through which data flows is achieved. A dynamic data sovereignty map is constructed to identify sensitive data storage locations and cross-border transmission paths. Topological network analysis is used to predict data flows that violate data localization requirements, providing a spatial dimension for intelligent interception decision-making.

[0048] (3) Adopt dynamic compliance strategy generation, apply graph neural networks to calculate the intersection of legal constraints in multiple jurisdictions, and establish a probabilistic model to evaluate legal applicability. Combined with real-time data flow topology, generate an optimization strategy to minimize compliance conflicts, dynamically adjust technical parameters such as data desensitization intensity and storage period, and resolve the problem of mutually exclusive legal requirements in cross-border scenarios;

[0049] (4) Adopt privacy computing sandbox execution and embed a compliance pre-verification engine in the federated learning framework. Achieve "available but invisible" data collaboration through secure multi-party computing protocols to ensure that the model training / prediction process complies with the target country's data export review standards;

[0050] (5) Adopt full-link audit tracing, use zero-knowledge proof technology to generate verifiable audit logs, build a distributed evidence chain, embed digital watermark technology to trace the source of the leak, and form a trusted accountability system throughout the entire cycle of data collection, transmission, calculation, and destruction.

[0051] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0053] Figure 1 This is a structural diagram of a cross-border data flow control system based on privacy computing verification.

[0054] Figure 2 It is a schematic diagram of the process of constructing the legal knowledge graph and multi-dimensional weight matrix of the present invention.

[0055] Figure 3 It is a flowchart of the present invention for updating the legal knowledge graph and the multi-dimensional weight matrix.

[0056] Figure 4 This is a schematic diagram of the pre-verification process of the privacy computing sandbox module of the present invention.

[0057] Figure 5 This is a flow chart of a cross-border data flow control method based on privacy computing verification in the present invention.

[0058] Figure 6 It is a schematic diagram of the structure of a computer-readable storage medium of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0060] Conventional cross-border data flow security management technologies in existing technologies are generally designed with a layered collaborative architecture as the core. In the data transmission link, an end-to-end encrypted channel is established through the SSL / TLS protocol or enterprise-level VPN to ensure that data is protected from eavesdropping and tampering during cross-border transmission; at the receiving end, it is based on a sovereign storage architecture and meets data residency requirements through geographic fencing and physical isolation; after the data is stored, the corresponding rule engine is used to automatically screen the data for compliance; and for data that needs to be shared across borders, dynamic desensitizing technology or homomorphic encryption technology is used to achieve confidential data transmission; finally, a technical trust foundation for multi-party collaboration is formed through the sovereignty declaration tag (DSV) and internationally recognized audit logs, and the further use of confidential computing in existing technologies is currently being integrated with the above technologies to promote the evolution of the entire cross-border data link from "static isolation" to "dynamic confidential collaboration".

[0061] However, the current mainstream cross-border data flow security management technologies, which mainly rely on technical means in their actual use, still have the following problems:

[0062] (1) Traditional encryption transmission technology uses channel encryption methods such as SSL / TLS and VPN to ensure transmission security, but the data still needs to be processed in plain text within the recipient's domain, which poses a risk of secondary leakage. That is, traditional encryption technology only ensures transmission security and cannot avoid plain text exposure during the data usage phase (such as cross-border joint modeling requires decrypted data); anonymization technology is susceptible to re-identification attacks (such as improper setting of differential privacy parameters leading to exhaustion of privacy budgets), and destroys data correlation, affecting analysis accuracy. The reason is that traditional technology separates the transmission, storage, and computing links, making it difficult to coordinate security and availability.

[0063] (2) Using localized data storage to meet data sovereignty requirements through physical isolation (such as domestic data centers) sacrifices data circulation efficiency and cannot support global business needs.

[0064] (3) Using static compliance review tools, compliance screening systems based on preset rule bases (such as keyword matching of GDPR clauses) rely on manual uploading of legal text updates, cannot respond to policy changes in real time, lack dynamic compliance adaptation capabilities, and are prone to legal-technical faults. At the same time, since compliance rules lack machine-parseable logical expressions, they cannot directly drive the automatic adaptation of technical systems, resulting in delayed updates of static compliance rule bases (such as the need for manual intervention to adjust sudden cross-border data blacklists), which easily leads to violation risks; legal clauses are out of touch with technical parameters (such as the lack of quantitative implementation standards for the "data minimization principle"), and rely on manual mapping based on expert experience.

[0065] (4) Anonymization / desensitization technology is used to mask and generalize sensitive fields, but it is difficult to balance privacy protection and data availability (for example, excessive desensitization of medical data leads to loss of analytical value).

[0066] (5) The lack of cross-border collaborative governance mechanisms makes it easy for sovereignty border conflicts to occur, leading to routing rigidity (such as the conflict between mandatory data localization and cross-border business needs) and lack of flexibility and elasticity; multi-party collaboration lacks a credible audit mechanism (such as the ambiguity of the responsibility for cross-border data transmission), making it difficult to meet the evidence requirements of multiple jurisdictions. Therefore, cross-border collaboration lacks a unified technical mutual trust foundation and dynamic arbitration mechanism, exacerbating sovereignty border conflicts.

[0067] The main technical problems mentioned above are the lag in compliance verification, the delayed response to policy changes resulting in compliance failure, the risk of legal sanctions and operational interruption for cross-border businesses, the rigidity of data flow paths restricting the efficiency of cross-border collaboration, increasing compliance costs, and easily causing conflicts over sovereign borders; secondly, there is the secondary problem of privacy leakage risk, that is, plaintext calculations leading to the exposure of sensitive data, triggering penalties under regulations such as GDPR and a crisis of user trust.

[0068] In response to the problems in the existing technology, the first embodiment of the present invention provides a cross-border data flow control system based on privacy computing verification to resolve the contradiction between the "risk of plaintext data leakage" and the "one-size-fits-all data localization storage" in traditional cross-border data flow. It realizes the "available but invisible" data through privacy computing technology (such as federated learning and multi-party secure computing), and constructs a "sovereign compliant digital twin" through blockchain evidence storage technology.

[0069] like Figure 1As shown, the embodiment of the present invention is a cross-border data flow control system based on privacy computing verification, including a legal knowledge graph construction module, a compliance policy dynamic generation module, a privacy computing sandbox module, a data flow topology analysis module and a full-link audit tracing module. Among them, the legal knowledge graph construction module is used to construct a legal knowledge graph and a multi-dimensional weight matrix of multi-national laws; the compliance policy dynamic generation module is used to generate a compliance policy combination based on the legal knowledge graph of multi-national laws; the privacy computing sandbox module is used to encrypt cross-border data and use the compliance policy combination in the distributed sandbox to pre-verify the encrypted cross-border data; the data flow topology analysis module is used to build a real-time topology network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multi-national laws into different topology nodes, and use the real-time topology network to control the flow of cross-border data; the full-link audit tracing module is used to monitor the data flow of cross-border data.

[0070] This invention realizes data compliance detection while effectively protecting data privacy by constructing a dynamic compliance verification mechanism of "privacy computing sandbox + dynamic legal engine + sovereignty boundary modeling" and "blockchain + privacy computing dual-engine architecture".

[0071] The following is a detailed description of the implementation effects of each module of this solution:

[0072] Legal knowledge graph construction module: In the embodiment of the present invention, the legal knowledge graph construction module is used to convert abstract legal clauses into executable technical parameters. Through the technical path of "legal digitization → conflict calculability → clause executableness", it realizes the qualitative change from text regulations to machine-readable rules, and supports the dynamic compliance adaptation of subsequent privacy computing.

[0073] S11. Legal digital mapping: preferably, Figure 2 As shown, the legal knowledge graph construction module first uses natural language processing (NLP) technology to parse the target country's legal text, extract key constraints, and generate structured legal metadata. Natural language processing technologies include BERT and GPT. Key constraints include the qualifications of the data recipient and the principle of minimizing transmission. Structured legal metadata records specific legal provisions. In this embodiment of the present invention, the legal knowledge graph is constructed using structured legal metadata as nodes and the mapping relationships between individual pieces of legal metadata as edges.

[0074] S12. Computable Conflicts: In particular, in the embodiments of the present invention, each node (legal clause) and edge (relationship between legal clauses) is assigned weight data. Each legal clause (node) may be assigned weights in multiple dimensions, such as time, jurisdictional conflict coefficient, etc. The setting of weights affects the priority of subsequent compliance decisions. To this end, the legal knowledge graph construction module in the embodiments of the present invention also establishes a multi-dimensional weight matrix to store the quantitative representation of the weights in the knowledge graph to support automated compliance decision-making and conflict resolution.

[0075] In this embodiment of the present invention, the conflict coefficients between various legal metadata and mapping relationships are determined from the legal knowledge graph, and a multidimensional weight matrix for each country's laws is constructed based on these conflict coefficients. The conflict coefficients include factors such as the time validity, scope of application, and conflict type of each node. Each conflict coefficient has a different weight, such as higher weights for nodes with newer timeframes and higher weights for clauses involving national security. Weights in the legal knowledge graph can be automatically adjusted through algorithms, such as decaying over time or dynamically changing based on the real-time status of the law.

[0076] In some embodiments, after constructing the legal knowledge graph and the multi-dimensional weight matrix, the legal knowledge graph construction module will also update them periodically.

[0077] S13. Scheduled Updates: Scheduled updates can be implemented through the incremental learning mechanism of the legal knowledge graph. First, regulatory revisions are tracked in real time. Upon discovery of regulatory updates, the logical relationships between nodes are dynamically adjusted to update the legal knowledge graph and the multi-dimensional weight matrix.

[0078] For example, Figure 3 As shown in Figure 2, the update of the legal knowledge graph and the multi-dimensional weight matrix specifically includes the following steps:

[0079] S131. Obtain the latest version of multinational legal text stream data: In an embodiment of the present invention, the legal knowledge graph construction module can obtain multinational legal texts through multinational government regulation publishing platforms (such as the judicial database of Country X) and international legal organization subscription interfaces (such as the World Trade Organization Committee, etc.). When a change in the regulation version number (such as "Country X Data Export Security Assessment Measures" v3.2→v4.0) or an update of the revision timestamp is detected, the revised text is automatically captured.

[0080] S132. Semantic Change Detection: In this embodiment of the present invention, the Legal Knowledge Graph Construction Module identifies semantic changes in legal text by comparing the semantic vectors of new and old clauses. Semantic changes include semantic additions (such as the addition of "the risk assessment period for data outbound transfer shall not exceed 12 months"), semantic deletions (such as the deletion of the original "cross-border transmission requires approval from a senior communications department"), and semantic shifts (such as changing the definition of "important data" from "100,000 people" to "involving critical security data").

[0081] S133. Knowledge Fusion and Conflict Resolution: When semantic changes are detected in newly acquired legal texts from multiple countries, the legal knowledge graph is reorganized and conflicts resolved to create a new legal knowledge graph. After the legal knowledge graph is updated, the multi-dimensional weight matrix can be updated based on the conflict coefficients of the new legal knowledge graph. For example, when a trade agreement between country X and country Y takes effect, the weights of related nodes and edges increase. Over time, the weights of nodes that have not been updated for a long time can also be automatically reduced.

[0082] During the knowledge graph update process, vertical and horizontal conflicts may arise due to the contradictions between various new and old laws. Vertical conflicts refer to conflicts between the metadata mapping relationships of the new and old laws of the same country after the reorganization (such as the differences between the new and old versions of the Personal Information Protection Law of Country X on the rights of data subjects); horizontal conflicts refer to conflicts between the metadata mapping relationships of laws of different countries after the reorganization (such as the different requirements for data localization between Country X and Country Y). At this time, the embodiment of the present invention adjusts the relationship between each node in the knowledge graph through conflict resolution means.

[0083] As a specific embodiment, conflict resolution specifically includes the following steps:

[0084] Detecting vertical and horizontal conflicts in the reorganized legal knowledge graph: Conflicts are represented by edges, which contain parameters such as conflict type, weight, and triggering conditions. For example, the conflict type can be vertical or horizontal, with the weight determined by the conflict coefficient and timeliness, and the triggering condition associated with changes in legal text version or jurisdiction.

[0085] According to the preset resolution strategy, vertical conflicts and horizontal conflicts are resolved: the resolution strategy set in the embodiment of the present invention includes soft resolution means and hard resolution means; the soft resolution means includes adjusting the confidence of the mapping relationship between legal metadata. Specifically, the confidence of the conflicting parties can be measured through a priority mechanism, and the legal interpretation of the party with higher confidence is selected to construct a logical relationship. The priority rules adopted by the soft resolution means include "new law prevails over old law" and "special law prevails over general law". Hard resolution means include establishing or deleting the mapping relationship between legal metadata. For example, when a vertical conflict is detected, the system will update the weight of the edge according to the new law, or directly disable the edge of the old law. The adjustment of the dynamic weight matrix is ​​also key. For example, changes in judicial weight and time limit weight will affect the processing of edges.

[0086] Conflicts that cannot be resolved (such as sovereignty-related clauses) will enter the manual review channel and be handled manually by professionals.

[0087] In cross-border data transfer scenarios, the legal knowledge graph construction module dynamically loads privacy regulations from multiple countries (such as the General Data Protection Regulation of Country X and the Personal Information Protection Law of Country Y), identifies legal conflicts through semantic analysis (such as the differences between data subject rights and outbound security assessment requirements), and automatically generates a legal conflict graph with priority rules marked (such as national security clauses taking precedence over commercial clauses).

[0088] This embodiment of the present invention utilizes a digital mapping of legal terms through a legal knowledge graph construction module. By analyzing privacy regulations across languages, it constructs a three-dimensional legal knowledge graph (jurisdiction / legal entity / obligation conditions), establishing a mapping relationship between legal terms and technical parameters. This enables multi-jurisdictional clause conflict detection, forms a machine-readable compliance benchmark rule base, and provides executable legal parameter input for subsequent modules.

[0089] Compliance policy dynamic generation module: In the embodiment of the present invention, the compliance policy dynamic generation module is used to convert legal clauses into executable technical policies and dynamically adapt to changes in cross-border data flows. Based on the real-time data flow topology and legal knowledge graph, it generates the intersection of legal constraints across jurisdictions to solve the policy optimization problem when the legal requirements of different jurisdictions conflict.

[0090] Specifically, in the compliance strategy dynamic generation module, generating a compliance strategy combination based on the legal knowledge graph of multiple countries' laws includes the following steps:

[0091] S21. Use graph neural networks to query key obligations of different national laws in the legal knowledge graph of multiple countries.

[0092] S22. Take the greatest common divisor of key obligations under the laws of different countries as the intersection of key obligations;

[0093] S23. Generate a combination of compliance strategies applicable to different countries based on the intersection of key obligations.

[0094] For example, regarding the deletion time limit for user outbound data, Country X requires deletion within 6 months, while Country Y requires deletion within 45 days. In this case, the greatest common denominator rule is 45-day deletion. Regarding the deletion scope of user data, Country X requires full-link deletion of outbound data, while Country Y requires simultaneous deletion of enterprise and third-party data. In this case, the greatest common denominator rule is full-link deletion. Regarding the verification of user requests, Country X requires verification of the requester's true identity, while Country Y requires verification of the requesting channel's identity. In this case, the greatest common denominator is dual verification involving both the requester and the requesting channel.

[0095] In the embodiment of the present invention, the compliance policy dynamic generation module adopts dynamic compliance policy generation, uses graph neural networks to calculate the intersection of legal constraints in multiple legal domains, and establishes a probabilistic model to evaluate legal applicability. Combined with the real-time data flow topology, it generates an optimization strategy for minimum compliance conflict, dynamically adjusts technical parameters such as data desensitization intensity and storage period, and solves the problem of mutual exclusion of legal requirements in cross-border scenarios. The compliance policy dynamic generation module outputs a compliance policy package containing a mapping table of technical parameters and legal terms (such as desensitization intensity = "appropriate security measures" requirements in Article 32 of Country Y's law, etc.). The compliance policy dynamic generation module outputs composite technical rules: for cross-border transportation data, it generates a policy matrix that integrates the requirements of multiple legal domains (such as a high-intensity encryption algorithm combined with a specific audit cycle), while avoiding restrictions on the processing of geographically sensitive information.

[0096] Privacy computing sandbox module: In the embodiment of the present invention, the privacy computing sandbox module realizes a distributed framework for cross-border data collaboration through federated learning, and the secure multi-party computing protocol realizes the "available but invisible" data throughout its life cycle. The compliance pre-verification engine dynamically loads the compliance requirements of the target country output by the legal knowledge graph, and implements automated compliance testing before data leaves the country.

[0097] Specifically, if Figure 4 As shown in Figure 1, the privacy computing sandbox module builds a distributed sandbox as a federated learning framework, embedding a compliance pre-verification engine within the federated learning framework to perform multi-party pre-verification between sandboxes. During the federated learning initialization phase, automated compliance verification and privacy enhancement are performed on the data of participating parties.

[0098] For horizontal federated learning scenarios (where the data features of the participants overlap but the samples are different), the engine performs differential privacy encryption ID alignment: first, it identifies sensitive identification fields (such as user mobile phone numbers and ID numbers) based on the legal knowledge graph, automatically calculates the privacy budget ε according to the privacy regulations of the target country (such as GDPR anonymization requirements), and injects Laplace noise into the identification field to generate a perturbed virtual ID; then, it uses the SHA-3 hash algorithm to irreversibly encrypt the virtual ID to ensure that each participant only exchanges ciphertext identifications during the sample alignment process, and the original identity information is invisible throughout the process.

[0099] For vertical federated learning (where participants have overlapping samples but different features), the engine uses RSA blind signature feature alignment: the data holder encrypts the local feature vector using the recipient's public key, generating a blinded feature request. The recipient blindly signs the encrypted feature (without obtaining the plaintext feature during the signing process) and returns the signature. The data holder verifies the signature, completing feature alignment. The entire process ensures the privacy of both feature selection permissions and feature values. The engine injects legal constraints in real time (e.g., Country X's mandatory desensitization of financial features) and dynamically adjusts encryption parameters (e.g., feature hash length and noise intensity) to ensure compliance and self-consistency throughout the federated training cycle.

[0100] The Privacy Computing Sandbox module utilizes a combination of oblivious transfer (OT) and homomorphic encryption (HE) protocols to ensure that plaintext data remains invisible during cross-border collaboration. The core logic of this combined protocol is that the OT layer protects the privacy of each party's choices during the computation process (e.g., data indexing and parameter selection); the HE layer protects the plaintext privacy of input data and supports encrypted computations (e.g., addition and multiplication).

[0101] Examples of the operational procedures for the combined protocol of oblivious transfer (OT) and homomorphic encryption (HE) include:

[0102] a. Data preparation: Each party encrypts local data using HE to generate an encrypted dataset.

[0103] b. Selective interaction: Through the OT protocol, the recipient secretly selects the encrypted data items or parameters that need to participate in the calculation.

[0104] c. Ciphertext computation: Based on the homomorphic properties of HE, joint computation is performed on encrypted data.

[0105] d. Result decryption: The final result is visible only to authorized parties.

[0106] Specifically, the distributed sandbox stores a set of compliance policies, which are used to jointly calculate encrypted gradient data. The results are then returned to the local cross-border data decryption server as pre-verification results. For example, if the pre-verification results contain sensitive transaction fields (e.g., ID number = high risk, transaction record = medium risk), desensitization rules (e.g., anonymization) are automatically triggered.

[0107] In this embodiment of the present invention, the privacy-preserving sandbox module supports cross-border joint modeling. When domestic and foreign institutions initiate requests through the sandbox, the pre-verification engine automatically intercepts original sensitive fields (such as user contact information) and forcibly replaces them with cryptographic hashes before initiating the federated learning process, ensuring that the original data never leaves the local domain. The privacy-preserving sandbox module enables "available but invisible" data collaboration through a secure multi-party computation protocol, ensuring that the model training and prediction processes comply with the target country's data export review standards.

[0108] Data flow topology analysis module: In the embodiment of the present invention, the data flow topology analysis module is used to build a real-time topology network for cross-border data transmission, identify the physical nodes and legal jurisdictions through which the data flows, and implement intelligent interception of data flows that violate data sovereignty or privacy protection principles.

[0109] S31. Execute real-time topology modeling: The data flow topology analysis module implements real-time topology modeling by constructing a digital twin of sovereign boundaries. Based on a graph database (such as Neo4j), this module dynamically maps physical nodes in the data flow path (e.g., cloud servers located in Country X) and their associated legal jurisdictions. Specifically, the data flow topology analysis module automatically constructs a virtualized sovereign compliance map by collecting IP geotags, cloud service provider node registration information, and blockchain-authenticated cross-border transmission logs along the data packet transmission path in real time. This map dynamically associates physical infrastructure (data centers, network nodes) with legal jurisdiction labels. For example, when data is transferred from Country X to Country Y via Country Z, the twin automatically labels the Country Z node as subject to the Country Z Personal Data Protection Act and calculates in real time the legal conflict coefficients for each jurisdiction through which the data flows (e.g., the conflict value between Country X's data localization requirements and Country Y's free flow principle).

[0110] S32. Establish an intelligent interception mechanism: The data flow topology analysis module pre-installs core data sovereignty clauses from multiple countries (such as Article 31 of Country X's Data Security Law, "Critical Information Infrastructure Data Localization") and converts them into programmable interception policies. When the real-time topology model detects data flowing through a transit node that has not passed security certification (such as a third-party server that has not received a security assessment), the rules engine immediately triggers a blocking action. The rule engine traverses the legal knowledge graph and multi-dimensional weight matrix involved in the data path to determine the overall compliance risk of the data path. If the overall compliance risk of the data path exceeds a preset threshold, the cross-border data is blocked or the path is switched.

[0111] The data flow topology analysis module of the embodiment of the present invention adopts data sovereignty boundary modeling, and tracks the physical nodes and jurisdictions through which data flows in real time based on IP geographic tags and blockchain transmission logs. The data flow topology analysis module optimizes the transmission path in real time: when data flows from a domestic server to an overseas recipient via a third-party transit node, the module automatically identifies the legal requirements of the transit area (such as data retention period restrictions), dynamically appends compliance policies or switches to a secure proxy node. The data flow topology analysis module also constructs a dynamic data sovereignty map, identifies sensitive data storage locations and cross-border transmission paths, and predicts data flows that violate data localization requirements through topological network analysis, providing a spatial dimension decision-making basis for intelligent interception.

[0112] Full-link Audit and Traceability Module: This module, implemented in this embodiment, embeds verifiable digital watermarks at each node in the data flow, enabling leak source tracking and accountability, breaking through the spatial and temporal barriers of cross-border data transmission and establishing a complete chain of causal evidence. Trusted digital watermarks, including hidden markers based on zero-knowledge proofs, embed irreversible watermarks at each node in the data flow, enabling real-time tracking of data flows.

[0113] During data flow, digital watermarks are dynamically superimposed along the transmission path: When each transit node receives data, the module hashes the current node information (IP address, jurisdiction code, etc.) to the previous watermark, generating a new composite watermark (e.g., Hash(previous watermark + node ID)), forming an irreversible chain of evidence. If data is leaked at an overseas recipient (e.g., illegally resold to a third party), the full-link audit and traceability module can extract the hash chain information from the final watermark through the watermark decoding engine. Combined with the sovereign boundary topology model stored in the graph database, it can accurately locate the leak (e.g., the decrypted log shows that the watermark's final node is the ID of a logistics provider from Country X, and the timestamp matches its data access record), enabling reconstruction of the leak link and the definition of responsibility.

[0114] Finally, the full-link audit traceability module writes encrypted logs of full-link watermark generation, compliance verification records (such as privacy sandbox pre-verification results), and data transmission path snapshots to consortium blockchain nodes (such as Hyperledger Fabric). Blockchain's distributed consensus mechanism ensures that audit evidence cannot be tampered with and supports multi-jurisdictional regulators to jointly verify data flow compliance through permissioned chain permissions (for example, a regulatory node in Country X can verify the implementation of Country X's clauses), forming a closed-loop audit system with multi-party governance.

[0115] The full-link audit tracing module of the embodiment of the present invention adopts full-link audit tracing, uses zero-knowledge proof technology to generate verifiable audit logs, builds a distributed evidence chain, embeds digital watermark technology to trace the source of the leak, and forms a trusted accountability system throughout the entire cycle of data collection, transmission, calculation, and destruction.

[0116] In summary, the embodiments of the present invention can achieve the following effects:

[0117] 1. The embodiment of the present invention uses the dynamic compliance verification mechanism of "privacy computing sandbox + dynamic legal engine + sovereignty boundary modeling" and "blockchain + privacy computing dual-engine architecture" to realize the construction of a cross-domain mutual trust foundation on the blockchain. Privacy computing ensures the confidential processing of data throughout its life cycle, resolves the contradiction between privacy protection and value mining, and can use practical legal knowledge graphs to convert natural language clauses into programmable compliance rules, drive the technical system to automatically adapt to policy changes, and realize the transition from single-point protection to full-link privacy computing, that is, covering the entire process of data encryption transmission, confidential computing, and compliance verification, and can analyze global laws and regulations in real time to generate executable technical strategies; from sovereignty confrontation to sovereignty collaboration: construct a sovereignty compliance map through digital twins to optimize data flow paths.

[0118] 2. The embodiments of the present invention are based on the dynamic balance between link privacy protection and data sovereignty compliance, and can resolve the contradiction between the "risk of plaintext data leakage" and the "one-size-fits-all data localization storage" in traditional cross-border data flows. Through privacy computing technologies (such as federated learning and multi-party secure computing), data can be made "available but invisible", and a "sovereign compliance digital twin" can be constructed through blockchain evidence storage technology.

[0119] 3. The embodiments of the present invention can solve the dilemma of "data silos" and "efficiency loss" in cross-border business scenarios by collaboratively optimizing the efficiency and security of cross-border data flow, and improve data circulation efficiency while satisfying data sovereignty boundary constraints.

[0120] 4. The embodiments of the present invention replace the traditional "post-compliance certification" model that relies on third-party audits by building a new trust mechanism for self-certification of technical compliance, thereby achieving "full-link verifiability and process traceability" compliance self-certification capabilities, which can reduce the risk of business interruption caused by regulatory conflicts for multinational companies.

[0121] 5. The embodiment of the present invention deeply couples privacy computing, dynamic legal analysis and sovereignty modeling through the triple innovation of "privacy computing sandbox + dynamic legal engine + sovereignty boundary modeling", forming a full-link closed loop.

[0122] 6. The embodiments of the present invention overcome the industry pain points of the inability to reconcile trusted records and privacy protection in cross-border data flows, compliance responses lagging behind policy changes, and business efficiency sacrificing security through the deep coupling of blockchain and privacy computing, dynamic mapping of law and technology, and spatial deduction of digital twins, thereby forming a verifiable, scalable, and auditable cross-border data infrastructure.

[0123] 7. The embodiments of the present invention, based on natural language processing (NLP) and legal rule extraction technology, convert abstract terms of privacy regulations into quantifiable and programmable technical parameters (such as data desensitization levels and minimum transmission thresholds). This can improve the efficiency of dynamic compliance verification, break the semantic gap between legal texts and algorithmic rules, and realize the automated mapping of "legal terms → logical rules → technical strategies", thus solving the lag problem of traditional compliance relying on manual translation.

[0124] 8. This embodiment of the present invention utilizes a federated learning framework and a secure multi-party computing protocol to achieve cross-border collaborative modeling without data leaving the country. It also embeds a compliance pre-verification engine to match the target country's legal constraints (such as the qualifications of the data recipient and storage period) in real time, significantly reducing the risk of privacy leaks. The privacy computing sandbox embeds a dynamic legal knowledge graph, further enabling the simultaneous updating of "technical execution" and "regulatory constraints," thereby avoiding compliance failures due to policy changes after data leaves the country.

[0125] 9. The embodiments of the present invention construct a data flow topology network, combine digital twin technology to simulate scenarios of overlapping legal jurisdictions of multiple countries, and optimize the compliance strategy matrix through game theory models (such as giving priority to the requirements of jurisdictions with higher privacy protection standards). This achieves cross-border collaboration cost optimization and a significant breakthrough in the balance between sovereignty and circulation, solves the problem of overlapping and conflicting data sovereignty boundaries of different countries, and dynamically generates transmission paths that minimize legal risks.

[0126] 10. The embodiment of the present invention is based on the blockchain's tamper-proof evidence storage mechanism and digital watermarking technology. By time-stamping all nodes of data flow (source, transmission path, and recipient), it can achieve traceability and definition of rights and responsibilities of leakage incidents, thereby further realizing the enhancement of technical self-certification compliance, breaking through the time-space barriers of cross-border data flow, and building a technical trust base for cross-judicial evidence recognition.

[0127] The second embodiment of the present invention discloses a cross-border data flow control method based on privacy computing verification, such as Figure 5 As shown, the following steps are included:

[0128] S41. Construct a legal knowledge graph and multi-dimensional weight matrix for laws in multiple countries;

[0129] S42. Generate compliance strategy combinations based on the legal knowledge graph of multiple countries’ laws;

[0130] S43. After encrypting the cross-border data, pre-verify the encrypted cross-border data using a compliance policy combination in the distributed sandbox;

[0131] S44. Build a real-time topological network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multiple countries' laws into different topological nodes, and use the real-time topological network to manage and control the flow of cross-border data;

[0132] S45. Monitor cross-border data flows.

[0133] The contents of the system in the first embodiment of the present invention are all applicable to the present method embodiment. The functions specifically implemented by the present method embodiment are the same as those of the above system embodiment, and the beneficial effects achieved are also the same as those achieved by the above system.

[0134] Figure 6 This is a structural diagram of the computer-readable storage medium of the third embodiment of the present invention. The computer-readable storage medium of the fourth embodiment of the present invention stores program instructions that can implement the above-mentioned cross-border data flow control system based on privacy computing verification, wherein the program instructions can be stored in the above-mentioned storage medium in the form of a software product, including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods of each embodiment of the present invention. The aforementioned computer-readable storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.

[0135] This embodiment also provides a computer program product. When the computer program product runs on a computer, it enables the computer to execute the above-mentioned related steps to implement the cross-border data flow control system based on privacy computing verification provided by the above embodiment.

[0136] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0137] Those skilled in the art will appreciate that the modules in the devices in the embodiments of the present invention can be adaptively changed and set in one or more devices different from the embodiments. The modules or units or components in the embodiments of the present invention can be combined into one module or unit or component, and in addition they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including corresponding claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner can be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including corresponding claims, abstracts and drawings) can be replaced by an alternative feature that provides the same, equivalent or similar purpose.

[0138] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0139] It should be noted that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0140] In addition, each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. In particular, for embodiments such as devices and equipment, since they are basically similar to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The embodiments of the devices and equipment described above are merely schematic, wherein the modules, units, etc. described as separate components may or may not be physically separated, that is, they may be located in one place, or they may be distributed to multiple places, such as nodes in a system network. Specifically, some or all of the modules and units may be selected according to actual needs to achieve the purpose of the above-mentioned embodiment scheme. Those skilled in the art can understand and implement it without paying any creative work.

[0141] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0142] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0143] In addition, the terms "first" and "second" used in the embodiments of the present invention are only used for descriptive purposes and should not be understood as indicating or implying relative importance, or implicitly indicating the number of technical features indicated in this embodiment. Therefore, the features defined by the terms "first" and "second" in the embodiments of the present invention can explicitly or implicitly indicate that the embodiment includes at least one of such features. In the description of the present invention, the word "plurality" means at least two or two or more, such as two, three, four, etc., unless otherwise clearly and specifically defined in the embodiments.

[0144] In the embodiments of the present invention, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus. In the absence of further restrictions, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or apparatus comprising the element. In addition, components, features, and elements with the same name in different embodiments of the present invention may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0145] Although embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention, and those of ordinary skill in the art may change, modify, replace, and modify the above embodiments within the scope of the present invention. Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the present invention. This application is intended to cover any variations, uses, or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary technical means in the art that are not disclosed in the present invention. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present invention are indicated by the claims below.

Claims

1. A cross-border data flow control system based on privacy computing verification, characterized by: It includes a legal knowledge graph construction module, a compliance policy dynamic generation module, a privacy computing sandbox module, a data flow topology analysis module, and a full-link audit tracing module; The legal knowledge graph construction module is used to construct a legal knowledge graph and a multi-dimensional weight matrix of laws of multiple countries; The compliance strategy dynamic generation module is used to generate a compliance strategy combination based on the legal knowledge graph of multiple countries' laws; The privacy computing sandbox module is used to encrypt cross-border data and then pre-verify the encrypted cross-border data using the compliance policy combination in the distributed sandbox; The data flow topology analysis module is used to build a real-time topology network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multiple countries' laws into different topology nodes, and use the real-time topology network to control the flow of cross-border data; The full-link audit and tracing module is used to monitor the data flow of cross-border data.

2. A cross-border data flow control system based on privacy computing verification according to claim 1, characterized in that: In the legal knowledge graph construction module, a legal knowledge graph and a multi-dimensional weight matrix of laws of multiple countries are constructed, specifically including the following steps: Obtain legal texts from multiple countries, generate structured legal metadata based on the legal texts, and map the relationship between each legal metadata to establish a legal knowledge graph; The conflict coefficients between the various legal metadata and mapping relationships are determined from the legal knowledge graph, and a multi-dimensional weight matrix of each country's laws is constructed based on the conflict coefficients.

3. A cross-border data flow control system based on privacy computing verification according to claim 2, characterized in that: The legal knowledge graph construction module is also used to regularly update the legal knowledge graph and the multi-dimensional weight matrix; the regularly updating the legal knowledge graph and the multi-dimensional weight matrix specifically includes the following steps: Regularly obtain the latest versions of legal texts from various countries; Detecting semantic changes in newly acquired legal texts from multiple countries; semantic changes include semantic addition, semantic reduction, and semantic shift; When semantic changes are found in newly acquired legal texts from multiple countries, the legal knowledge graph is reorganized and conflicts are resolved to obtain a new legal knowledge graph; The multi-dimensional weight matrix is ​​updated according to the conflict coefficient of the new legal knowledge graph.

4. A cross-border data flow control system based on privacy computing verification according to claim 3, characterized in that: The conflict resolution specifically includes the following steps: Detecting vertical and horizontal conflicts in the reorganized legal knowledge graph; the vertical conflict represents the conflict between the mapping relationships of new and old legal metadata of the same country after the reorganization; the horizontal conflict represents the conflict between the mapping relationships of legal metadata of different countries after the reorganization; Resolving the vertical and horizontal conflicts according to a preset resolution strategy; the resolution strategy includes soft resolution and hard resolution; the soft resolution includes adjusting the confidence level of the mapping relationship between legal metadata, and the hard resolution includes establishing or deleting the mapping relationship between legal metadata; For conflicts that cannot be resolved by the preset resolution strategy, manual review is introduced.

5. The cross-border data flow control system based on privacy computing verification according to claim 1 is characterized in that: In the compliance strategy dynamic generation module, generating a compliance strategy combination based on the legal knowledge graph of multiple countries' laws specifically includes the following steps: Query key obligations of different countries’ laws through graph neural networks in the legal knowledge graph of multiple countries’ laws; Take the greatest common divisor of key obligations under the laws of different countries as the intersection of key obligations; Generate a portfolio of compliance strategies suitable for different countries based on the intersection of key obligations.

6. A cross-border data flow control system based on privacy computing verification according to claim 1, characterized in that: In the privacy computing sandbox module, cross-border data is locally encrypted through the OT-HE combined protocol, and the encrypted gradient data is transmitted between distributed sandboxes; The distributed sandbox stores a compliance policy combination, which is used to jointly calculate the encrypted gradient data, and returns the calculation result to the local decryption of the cross-border data as a pre-verification result.

7. The cross-border data flow control system based on privacy computing verification according to claim 1 is characterized in that: In the data flow topology analysis module, the real-time topology network is used to control the flow of cross-border data, specifically including the following steps: Determine the data path of cross-border data in the topological network, traverse the legal knowledge graph and multi-dimensional weight matrix involved in the data path, and determine the overall compliance risk of the data path; When the total compliance risk of the data path exceeds the preset threshold, cross-border data will be blocked or the path will be switched.

8. The cross-border data flow control system based on privacy computing verification according to claim 1 is characterized in that: The full-link audit tracing module is used to add digital watermarks to each data processing node of the legal knowledge graph construction module, the compliance policy dynamic generation module, the privacy computing sandbox module and the data flow topology analysis module, and monitor and store the data flow of cross-border data through digital watermarks; and write the stored content into the blockchain.

9. A cross-border data flow control method based on privacy computing verification, characterized in that: The following steps are included: Construct legal knowledge graphs and multi-dimensional weight matrices of laws from multiple countries; Generate compliance strategy combinations based on the legal knowledge graph of multiple countries’ laws; After encrypting the cross-border data, pre-verify the encrypted cross-border data using the compliance policy combination in the distributed sandbox; Construct a real-time topological network for cross-border data transmission, load the legal knowledge graph and multi-dimensional weight matrix of multiple countries' laws into different topological nodes, and use the real-time topological network to control the flow of cross-border data; Monitor data flows across borders.

10. A computer-readable storage medium, characterized in that The storage medium stores a program for implementing the system according to any one of claims 1 to 8, and the program can be executed by a processor.

Images