Multi-dimensional encryption transmission protection method and system for cross-border payment

Abstract

The present application relates to a multi-dimensional encryption transmission protection method and system for cross-border payment, belonging to the field of electronic digital data processing, more specifically to the field of Internet security services, the method comprising: using a gradient boosting decision tree and deep cross network fusion model to intelligently analyze a dynamic multi-dimensional encryption strategy uniquely corresponding to a current cross-border payment transaction according to each item of basic data associated with the current cross-border payment transaction. The present application also relates to a multi-dimensional encryption transmission protection system for cross-border payment. Through the present application, in view of the technical problems in the prior art that the encryption transmission protection strategies for each cross-border payment are difficult to balance security, transaction efficiency and power saving, and the corresponding key protection capability is insufficient, a gradient boosting decision tree and deep cross network fusion model designed with a customized structure is used to intelligently analyze a unique dynamic multi-dimensional encryption strategy for each cross-border payment transaction according to each item of basic data selected in a targeted manner, thereby solving the above technical problems.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing, and more specifically to the field of internet security services, particularly to a multi-dimensional encrypted transmission protection method and system for cross-border payments. Background Technology

[0002] Digital data processing can be specifically applied to internet security services to achieve security control in various internet security application scenarios, such as encrypted transmission protection for cross-border payments. Clearly, cross-border payments involve significant payment risks, requiring key encryption for each cross-border payment request carrying various payment data before the encrypted data can be transmitted over the internet. Generally, without considering the specific payment scenario, specific payment data, specific information of the paying parties, or specific purchase data for each cross-border payment, a similar encrypted transmission protection strategy is designed for all cross-border payments. This encryption transmission protection strategy has a high level of protection, thus simultaneously meeting the risk control needs of each cross-border payment.

[0003] For example, Chinese invention patent publication CN114157468A proposes a method and apparatus for transmitting cross-border payment messages, applicable to the field of cross-border payment technology. The method includes: a standard transceiver corresponding to an indirect participant system receiving a first message stored in a local queue by the indirect participant system; parsing the first message to determine first initiator information and first recipient information; obtaining a first processing component list based on the first initiator information, first recipient information, and uplink processing type; encrypting the first message after rule verification based on the first processing component list information; and sending the encrypted first message to the standard transceiver corresponding to the direct participant system according to the remote queue corresponding to the first message routing domain information. This invention can ensure the standardization of CIPS standard message transmission and achieve the independence, integrity, and security of standard transmission in cross-border RMB business.

[0004] For example, Chinese invention patent publication CN120710759A proposes a method for encrypted transmission and storage of financial data. Its key features include: sensitivity-driven data hierarchical fragmentation at the user terminal; dynamic elliptic curve encryption for transmitted data and fully homomorphic encryption for stored data using a dual-channel encryption engine; dynamic allocation of fragments to heterogeneous cloud nodes based on reinforcement learning-based multi-cloud routing; and the construction of a distributed key management matrix to disperse key fragments across the blockchain and hardware security modules, with reconstruction activated by biometric features. The advantages of this method include achieving end-to-end encrypted operations, eliminating the risk of plaintext exposure; supporting encrypted state financial computation; resistance to single points of failure; dynamic defense against APT attacks; quantum security evolution capabilities; and applicability to scenarios such as mobile banking and cross-border payments.

[0005] Therefore, it is evident that the encryption transmission protection strategies for various cross-border payments in existing technologies are basically the same, failing to consider the specific payment scenario, various payment data, specific information of the paying parties, and specific data of the purchased goods for each cross-border payment. Furthermore, designing the same encryption transmission protection strategy for each cross-border payment makes it difficult to balance security, transaction efficiency, and computational resource conservation. For example, if the encryption transmission protection strategy is designed with a high protection level, it is difficult to guarantee the transaction efficiency and computational resource conservation of each cross-border payment. Conversely, if the encryption transmission protection strategy is designed with a low protection level, the security of each cross-border payment cannot be guaranteed. At the same time, existing encryption transmission protection strategies for each cross-border payment cannot comprehensively consider five dimensions, including encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. This results in an overly simplistic encryption transmission protection strategy for each cross-border payment, making it difficult to effectively reduce the probability of key cracking. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides a multi-dimensional encrypted transmission protection method and system for cross-border payments. By leveraging a gradient boosting decision tree fusion model with a customized structure and a deep cross-network, and based on selectively chosen fundamental data, a unique dynamic multi-dimensional encryption strategy is intelligently analyzed for each cross-border payment transaction. This dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. This employs an artificial intelligence model to customize different dynamic multi-dimensional encryption strategies for each cross-border payment transaction, achieving a balance between security, transaction efficiency, and computational resource conservation. Furthermore, the simultaneous introduction of the five dimensions in the encrypted transmission protection strategy further reduces the probability of key cracking.

[0007] According to a first aspect of the present invention, a multi-dimensional encrypted transmission protection method for cross-border payments is provided, the method comprising:

[0008] For each cross-border payment transaction, the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties are collected in real time to output the payment-related content of the cross-border payment transaction.

[0009] For each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to output multiple sets of product sales data for that cross-border payment transaction.

[0010] The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier.

[0011] The payment request for the current cross-border payment transaction is encrypted, signed, and encapsulated according to the dynamic multi-dimensional encryption strategy to form a payment message. The payment message and the policy digest of the dynamic multi-dimensional encryption strategy are transmitted concurrently to the recipient's transaction device through multiple channels.

[0012] The recipient's transaction device reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy to complete the decryption operation of the payment message.

[0013] According to a second aspect of the present invention, a multi-dimensional encrypted transmission protection system for cross-border payments is provided. The system includes a memory and a plurality of processors. The memory stores a computer program configured to be executed by the plurality of processors to complete the following steps:

[0014] For each cross-border payment transaction, the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties are collected in real time to output the payment-related content of the cross-border payment transaction.

[0015] For each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to output multiple sets of product sales data for that cross-border payment transaction.

[0016] The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier.

[0017] The payment request for the current cross-border payment transaction is encrypted, signed, and encapsulated according to the dynamic multi-dimensional encryption strategy to form a payment message. The payment message and the policy digest of the dynamic multi-dimensional encryption strategy are transmitted concurrently to the recipient's transaction device through multiple channels.

[0018] The payee's transaction device reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy to complete the decryption operation of the payment message. This step should be implemented on the payee's transaction device side, but in this system, it can be simulated by the processor on the local side, i.e., the payer's transaction device side.

[0019] According to a third aspect of the present invention, a multi-dimensional encrypted transmission protection system for cross-border payments is provided, the system comprising:

[0020] The first capturing device is used to collect in real time the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score value of both parties, and the sanction status identifier of both parties for each cross-border payment transaction, so as to output each payment-related content of the cross-border payment transaction.

[0021] The second capturing device is used to collect in real time the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction for each cross-border payment transaction, so as to output multiple product sales data for that cross-border payment transaction.

[0022] The intelligent parsing device, connected to the first and second capturing devices, is used to intelligently parse the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction, and multiple commodity sales data using a gradient boosting decision tree and deep cross-network fusion model. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0023] The message generation device, connected to the intelligent parsing device, is used to encrypt, sign, and encapsulate the payment request of the current cross-border payment transaction according to the dynamic multi-dimensional encryption strategy to form a payment message, and transmits the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy concurrently to the recipient's transaction device through multiple channels.

[0024] The decryption processing device, located on the recipient's transaction device and connected to the message generation device, is used to reconstruct the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy in order to complete the decryption operation of the payment message.

[0025] Compared with the prior art, the present invention has at least the following outstanding substantive features:

[0026] (1) The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy unique to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data of the current cross-border payment transaction. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier. Thus, artificial intelligence mode is used to customize different dynamic multi-dimensional encryption strategies for each cross-border payment transaction. While ensuring the protection effect of multi-dimensional encryption transmission for cross-border payment, it also takes into account the transaction efficiency improvement and computing resource saving of each cross-border payment transaction.

[0027] (2) A customized artificial intelligence model is introduced for the current cross-border payment transaction intelligent parsing unique dynamic multi-dimensional encryption strategy. The artificial intelligence model is a gradient boosting decision tree and deep cross network fusion model. The number of training times performed on the gradient boosting decision tree and deep cross network fusion model is the same as the numerical change trend of the total number of transaction device types. The customized structure design of the above artificial intelligence model ensures the stability and reliability of the intelligent parsing results of the dynamic multi-dimensional encryption strategy.

[0028] (3) In each training of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the gradient boosting decision tree and deep cross-network fusion model, and the payment scenario type number, payment amount, payment association content of the historical cross-border payment transaction and multiple commodity sales data are used as the input data of the deep cross-network fusion model to complete the training, thereby ensuring the training effect of each training.

[0029] (4) Targeted screening of various basic data is introduced for the current cross-border payment transaction intelligent analysis uniquely corresponding dynamic multi-dimensional encryption strategy. The basic data includes the payment scenario type number of the current cross-border payment transaction, the payment amount, each payment-related content of the current cross-border payment transaction, and multiple commodity sales data of the current cross-border payment transaction. Each payment-related content of the current cross-border payment transaction is the transaction device type number of the two parties in the current cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties. The multiple commodity sales data of the current cross-border payment transaction is the commodity type number of the purchased commodity in the current cross-border payment transaction, the historical total sales, the historical average sales amount, and the transaction-related information of each historical cross-border payment transaction. The transaction-related information of each historical cross-border payment transaction for purchasing the commodity is the payment-related content of the historical cross-border payment transaction for purchasing the commodity. The targeted screening of the above basic data further ensures the stability and reliability of the intelligent analysis results of the dynamic multi-dimensional encryption strategy.

[0030] (5) For each cross-border payment transaction, the number of historical cross-border payment transactions for purchasing goods is positively correlated with the total number of existing transaction device types. Each historical cross-border payment transaction is the closest historical cross-border payment transaction to the current cross-border payment transaction for purchasing the goods. A numerical conversion formula is used to represent the numerical conversion relationship between the number of historical cross-border payment transactions and the total number of existing transaction device types, thereby completing the customized design of the data structure for the basic data used for intelligent parsing. Attached Figure Description

[0031] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0032] Figure 1 This is a schematic diagram of the working scenario of the multi-dimensional encrypted transmission protection method and system for cross-border payments according to the present invention.

[0033] Figure 2 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 1 of the present invention.

[0034] Figure 3 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 2 of the present invention.

[0035] Figure 4 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 3 of the present invention.

[0036] Figure 5This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 4 of the present invention.

[0037] Figure 6 This is a schematic diagram of a multi-dimensional encrypted transmission protection system for cross-border payments, as shown in Embodiment 5 of the present invention.

[0038] Figure 7 This is a schematic diagram of a multi-dimensional encrypted transmission protection system for cross-border payments, as shown in Embodiment 6 of the present invention. Detailed Implementation

[0039] like Figure 1 The diagram illustrates a working scenario of a multi-dimensional encrypted transmission protection method and system for cross-border payments, based on the present invention. This invention relates to the field of electronic digital data processing, and more specifically to the field of internet security services.

[0040] The specific technical process of this invention is as follows:

[0041] Technical Process A: To intelligently analyze the different dynamic multi-dimensional encryption strategies corresponding to each cross-border payment transaction, a custom-designed artificial intelligence model is introduced, such as... Figure 1 As shown;

[0042] like Figure 1 As shown, the current payment process for cross-border payment transactions is illustrated. Figure 1 As shown in the box on the left, the current cross-border payment transaction is a cross-border payment transaction in which an overseas payer sends a payment request to a payee in China. Of course, the cross-border payment transactions involved in this invention are not limited to those described above. Figure 1 The payment process is shown below;

[0043] Specifically, the customized structural design of the introduced artificial intelligence model is mainly reflected in the following aspects:

[0044] Firstly, the artificial intelligence model described is a fusion model of gradient boosting decision tree and deep cross-network;

[0045] Secondly, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model shows the same trend as the total number of existing transaction device types.

[0046] Thirdly: In each training iteration of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the gradient boosting decision tree and deep cross-network fusion model, while the payment scenario type number, payment amount, payment-related content of the historical cross-border payment transaction, and multiple commodity sales data are used as the input data of the deep cross-network fusion model to complete the training, thereby ensuring the training effect of each iteration.

[0047] In this way, the stability and reliability of the intelligent parsing results of the dynamic multi-dimensional encryption strategy are ensured through the customized structural design of the above artificial intelligence model.

[0048] Technical Process B: To execute intelligent parsing of different dynamic multi-dimensional encryption strategies corresponding to each cross-border payment transaction, various basic data are selectively filtered.

[0049] like Figure 1 As shown, the basic data includes each payment-related item of the current cross-border payment transaction, multiple commodity sales data of the current cross-border payment transaction, and other transaction-related information of the current cross-border payment transaction. The other transaction-related information of the current cross-border payment transaction includes the payment scenario type number and payment amount of the current cross-border payment transaction.

[0050] Specifically, the basic data includes the payment scenario type number of the current cross-border payment transaction, the payment amount, each payment-related content of the current cross-border payment transaction, and multiple sets of product sales data for the current cross-border payment transaction. Each payment-related content of the current cross-border payment transaction includes the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status identifier of both parties. The multiple sets of product sales data for the current cross-border payment transaction include the product type number of the product purchased in the current cross-border payment transaction, the historical total sales, the historical average sales amount, and the transaction-related information of each historical cross-border payment transaction. The transaction-related information of each historical cross-border payment transaction for the purchase of the product is the payment-related content of that historical cross-border payment transaction for the purchase of the product.

[0051] More specifically, for each cross-border payment transaction, the number of historical cross-border payment transactions for purchasing goods is positively correlated with the total number of existing transaction device types. Each historical cross-border payment transaction is the closest historical cross-border payment transaction to the current cross-border payment transaction for purchasing the goods. A numerical conversion formula is used to represent the numerical conversion relationship between the number of historical cross-border payment transactions and the total number of existing transaction device types, thereby completing the customized design of the data structure for the basic data used for intelligent parsing.

[0052] In this way, the stability and reliability of the intelligent parsing results of the dynamic multi-dimensional encryption strategy are further guaranteed through the targeted screening of the above basic data.

[0053] Technical Process C: Using the AI ​​model with a customized structure designed based on Technical Process A, and based on the basic data specifically selected by Technical Process B, the model intelligently analyzes different dynamic multi-dimensional encryption strategies for each cross-border payment transaction.

[0054] Specifically, the dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. The customization of the multi-dimensional encryption strategy improves the protection effect of encrypted transmission for cross-border payments.

[0055] Specifically, such as Figure 1 As shown, for the current cross-border payment transaction, a dynamic multi-dimensional encryption strategy is provided that is uniquely corresponding to the intelligent parsing of the current cross-border payment transaction. The dynamic multi-dimensional encryption strategy output by the intelligent parsing provides specific values ​​for each of the following: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0056] Technical Process D: Based on Technical Process C, the payment request for each cross-border payment transaction is encrypted, signed, and encapsulated using a unique dynamic multi-dimensional encryption strategy that is intelligently parsed to form a payment message. The payment message and its accompanying strategy digest are then transmitted concurrently to the payee's transaction device through multiple channels.

[0057] Specifically, for the current cross-border payment transaction, the payment request of the current cross-border payment transaction is encrypted, signed and encapsulated based on the unique corresponding dynamic multi-dimensional encryption strategy of the current cross-border payment transaction intelligent parsing to form a payment message, and the payment message and the policy digest of the attached dynamic multi-dimensional encryption strategy are concurrently transmitted to the payee's transaction device through multiple channels;

[0058] Technical Process E: For each cross-border payment transaction, the recipient's transaction device reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy uniquely corresponding to the cross-border payment transaction to complete the decryption operation of the payment message;

[0059] Specifically, for the current cross-border payment transaction, the recipient's transaction device reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction to complete the decryption operation of the payment message;

[0060] Therefore, through the coordinated operation of the above five technical processes, a dynamic multi-dimensional encryption strategy uniquely corresponding to each cross-border payment transaction is achieved by using a gradient boosting decision tree and deep cross-network fusion model. This model intelligently analyzes the payment scenario type number, payment amount, related payment content of the current cross-border payment transaction, and multiple sales data of goods in the current cross-border payment transaction. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. Thus, different dynamic multi-dimensional encryption strategies are customized for each cross-border payment transaction using an artificial intelligence model. This ensures the multi-dimensional encryption transmission protection effect for cross-border payments while also improving transaction efficiency and saving computing resources.

[0061] The key points of this invention are: intelligent parsing of different dynamic multi-dimensional encryption strategies corresponding to each cross-border payment transaction; customized structural design of the artificial intelligence model used for intelligent parsing; introduction of multi-dimensional encryption strategies including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier; and targeted screening of various basic data used for intelligent parsing.

[0062] The multi-dimensional encrypted transmission protection method and system for cross-border payments of the present invention will be specifically described below by way of embodiments.

[0063] Example 1

[0064] Figure 2 The following is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 1 of the present invention. The method can be applied to a payment management server that performs cross-border payment management.

[0065] like Figure 2 As shown, the multi-dimensional encrypted transmission protection method for cross-border payments includes the following specific steps:

[0066] Step S1: For each cross-border payment transaction, collect the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties in real time to output the payment-related content of the cross-border payment transaction.

[0067] Specifically, the transaction devices of both parties in each cross-border payment transaction can be tablets, laptops, mobile phones, desktops, servers, and other transaction device types, with different transaction device types corresponding to different transaction device type numbers;

[0068] Specifically, the IP geographic information of both parties in each cross-border payment transaction can be the geographic name of the region to which the IP addresses currently used by both parties belong. For example, it can be the ASCII code value corresponding to the geographic name of the region to which the IP addresses currently used by both parties belong.

[0069] Step S2: For each cross-border payment transaction, collect in real time the product type number of the purchased goods, the total historical sales volume, the historical average sales amount, and transaction-related information of each historical cross-border payment transaction to output multiple sets of product sales data for that cross-border payment transaction.

[0070] Specifically, for each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to serve as multiple product sales data outputs for that cross-border payment transaction. The products purchased for each cross-border payment transaction can be of various types, such as electronic facilities, luxury goods, and office supplies.

[0071] Step S3: Using a gradient boosting decision tree and deep cross-network fusion model, the current cross-border payment transaction's payment scenario type number, payment amount, various payment-related contents of the current cross-border payment transaction, and multiple commodity sales data are used to intelligently parse the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0072] Specifically, the intelligent parsing provides five specific values ​​for the five dimensions of the current cross-border payment transaction's unique dynamic multi-dimensional encryption strategy: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0073] Step S4: Encrypt, sign, and encapsulate the payment request of the current cross-border payment transaction according to the dynamic multi-dimensional encryption strategy to form a payment message, and transmit the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy concurrently to the recipient's transaction device through multiple channels;

[0074] Specifically, the payment request of the current cross-border payment transaction is encrypted, signed, and encapsulated to form a payment message based on a dynamic multi-dimensional encryption strategy, and the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy are concurrently transmitted to the recipient's transaction device through multiple channels. The channel isolation level is used to securely isolate the multiple channels. The higher the channel isolation level, the more stringent the security isolation of the multiple channels, that is, the higher the encryption degree of the corresponding dynamic multi-dimensional encryption strategy.

[0075] Step S5: Reconstruct the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy to complete the decryption operation of the payment message;

[0076] Among the five dimensions of dynamic multidimensional encryption strategy, including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multidimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multidimensional encryption strategy.

[0077] For example, the quantum key distribution identifier used to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy includes: using different binary values ​​to represent the quantum key distribution identifier in different ways to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy;

[0078] Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list.

[0079] For example, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party to the payment, and the sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list, including: using different binary values ​​to represent the sanction status identifier of each party to the payment to indicate whether the party is on the cross-border transaction sanction list.

[0080] For example, different payment scenario types correspond to different payment scenario type numbers. The two parties to the payment are the payer and the payee. The historical fraud score value of each party to the payment is the percentage of cross-border payment transactions that the party has participated in that were rated as fraudulent by the other party. The sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. It also includes: the cross-border transaction sanction list has an overflow limit on the maximum number of sanctions.

[0081] Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties.

[0082] In each training iteration of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the model, while the payment scenario type number, payment amount, payment-related content of the transaction, and multiple sales data are used as the input data of the deep cross-network fusion model to complete the training.

[0083] Specifically, numerical simulation mode can be used to test and simulate each training iteration of the gradient boosting decision tree and deep cross-network fusion model.

[0084] Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types;

[0085] For example, the numerical trends of the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model and the total number of existing transaction device types are the same, including: when there are 50 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1000; when there are 60 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1200; when there are 70 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1400; when there are 80 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1600, and so on.

[0086] Example 2

[0087] Figure 3 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 2 of the present invention.

[0088] like Figure 3 As shown, with Figure 2 Unlike the previous implementation, after employing a gradient boosting decision tree and deep cross-network fusion model to intelligently parse the current cross-border payment transaction's unique dynamic multi-dimensional encryption strategy based on the payment scenario type number, payment amount, various payment-related contents of the current cross-border payment transaction, and multiple commodity sales data, the dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier, i.e., after step S3, the method further includes:

[0089] Step S6: Receive the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction, and store the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction on the payer's transaction device.

[0090] Among them, receiving the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction and storing the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction on the payer's transaction device includes: using a dynamic storage chip to complete the storage of the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction.

[0091] Alternatively, FLASH flash memory can be used to replace the dynamic storage chip for storing the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction;

[0092] The process of receiving the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction and storing the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction on the payer's transaction device also includes setting a storage period for the storage of the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction.

[0093] Example 3

[0094] Figure 4 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 3 of the present invention.

[0095] like Figure 4 As shown, with Figure 2Unlike the previous implementation, before real-time collection of the transaction device type numbers of both parties, their IP geographic information, historical fraud scores, and sanction status identifiers for each cross-border payment transaction to output the associated payment content for that transaction, i.e., before step S1, the method further includes:

[0096] Step S7: Perform multiple training iterations on the gradient boosting decision tree and deep cross-network fusion model. The more types of transaction devices there are, the more times the gradient boosting decision tree and deep cross-network fusion model should be trained.

[0097] For example, a numerical transformation formula can be used to represent the numerical transformation relationship between the total number of existing transaction device types and the number of training operations performed on the gradient boosting decision tree and deep cross-network fusion model.

[0098] Example 4

[0099] Figure 5 This is a flowchart illustrating the steps of a multi-dimensional encrypted transmission protection method for cross-border payments according to Embodiment 4 of the present invention.

[0100] like Figure 5 As shown, with Figure 2 Unlike the previous implementation, after employing a gradient boosting decision tree and deep cross-network fusion model to intelligently parse the current cross-border payment transaction's unique dynamic multi-dimensional encryption strategy based on the payment scenario type number, payment amount, various payment-related contents of the current cross-border payment transaction, and multiple commodity sales data, the dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier, i.e., after step S3, the method further includes:

[0101] Step S8: Wirelessly transmit the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction to the remote encryption management server;

[0102] For example, wirelessly transmitting the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction to the remote encryption management server includes: the remote encryption management server being a big data service node or a blockchain service node, and the encryption management server and the payment management server being located in different positions on the network;

[0103] The wireless transmission of the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction to the remote encryption management server includes: the wireless transmission being encrypted transmission.

[0104] Next, the various method embodiments of the present invention will be described in detail.

[0105] In the multi-dimensional encrypted transmission protection method for cross-border payments according to various method embodiments of the present invention:

[0106] In a dynamic multi-dimensional encryption strategy that includes five dimensions such as encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier, quantum key distribution utilizes the properties of quantum mechanics to generate and share a random key for both parties in a cross-border payment transaction.

[0107] Specifically, in the dynamic multi-dimensional encryption strategy that includes five dimensions such as encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier, quantum key distribution is to use the properties of quantum mechanics to generate and share a random key for both parties in a cross-border payment transaction. This includes: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier all acting simultaneously on the generation and transmission of a random key.

[0108] Quantum key distribution utilizes quantum mechanical properties to generate and share a random key for both parties in a cross-border payment transaction. Specifically, in quantum key distribution, if a third party outside the two parties attempts to eavesdrop on the key, the cryptographic analysis operation used to eavesdrop will generate detectable anomalies. These anomalies are transmitted through quantum superposition or quantum entanglement states, allowing the two parties in the cross-border payment transaction to detect the presence of eavesdropping. When the number of third parties eavesdropping on the generated random key is less than or equal to a set threshold, the generated random key is shared with both parties in the cross-border payment transaction.

[0109] And in the multi-dimensional encrypted transmission protection method for cross-border payments according to various method embodiments of the present invention:

[0110] The numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model is the same as that of the total number of existing transaction device types, including: using a total number change curve to represent the numerical trend of the total number of existing transaction device types, and using a number change curve to represent the numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model.

[0111] Among them, the numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model being the same as the total number of existing transaction device types also includes: the curvature values ​​at uniform intervals on the total number change curve are equal to the curvature values ​​at uniform intervals on the number of training sessions change curve.

[0112] Specifically, the curvature values ​​at uniformly spaced locations on the total number variation curve are equal to the curvature values ​​at uniformly spaced locations on the number variation curve, including: the number of uniformly spaced locations on the total number variation curve is equal to the number of uniformly spaced locations on the number variation curve.

[0113] Furthermore, the fact that the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model shows the same numerical trend as the total number of existing transaction device types also includes: using programmable logic devices to simulate and model the total number change curve and the number of iterations change curve.

[0114] Example 5

[0115] Figure 6 This is a schematic diagram of a multi-dimensional encrypted transmission protection system for cross-border payments, as shown in Embodiment 5 of the present invention. The system can be applied to a payment management server that performs cross-border payment management.

[0116] like Figure 6 As shown, the multi-dimensional encrypted transmission protection system for cross-border payments includes a memory and multiple processors. The memory stores a computer program configured to be executed by the multiple processors to complete the following steps:

[0117] For each cross-border payment transaction, the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties are collected in real time to output the payment-related content of the cross-border payment transaction.

[0118] Specifically, the transaction devices of both parties in each cross-border payment transaction can be tablets, laptops, mobile phones, desktops, servers, and other transaction device types, with different transaction device types corresponding to different transaction device type numbers;

[0119] Specifically, the IP geographic information of both parties in each cross-border payment transaction can be the geographic name of the region to which the IP addresses currently used by both parties belong. For example, it can be the ASCII code value corresponding to the geographic name of the region to which the IP addresses currently used by both parties belong.

[0120] For each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to output multiple sets of product sales data for that cross-border payment transaction.

[0121] Specifically, for each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to serve as multiple product sales data outputs for that cross-border payment transaction. The products purchased for each cross-border payment transaction can be of various types, such as electronic facilities, luxury goods, and office supplies.

[0122] The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier.

[0123] Specifically, the intelligent parsing provides five specific values ​​for the five dimensions of the current cross-border payment transaction's unique dynamic multi-dimensional encryption strategy: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0124] The payment request for the current cross-border payment transaction is encrypted, signed, and encapsulated according to the dynamic multi-dimensional encryption strategy to form a payment message. The payment message and the policy digest of the dynamic multi-dimensional encryption strategy are transmitted concurrently to the recipient's transaction device through multiple channels.

[0125] Specifically, the payment request of the current cross-border payment transaction is encrypted, signed, and encapsulated to form a payment message based on a dynamic multi-dimensional encryption strategy, and the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy are concurrently transmitted to the recipient's transaction device through multiple channels. The channel isolation level is used to securely isolate the multiple channels. The higher the channel isolation level, the more stringent the security isolation of the multiple channels, that is, the higher the encryption degree of the corresponding dynamic multi-dimensional encryption strategy.

[0126] The payment message is decrypted by reconstructing the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy on the recipient's transaction device. This step should be implemented on the recipient's transaction device. In this embodiment, it can be simulated by the processor on the local end, i.e., the payer's transaction device.

[0127] Among the five dimensions of dynamic multidimensional encryption strategy, including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multidimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multidimensional encryption strategy.

[0128] For example, the quantum key distribution identifier used to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy includes: using different binary values ​​to represent the quantum key distribution identifier in different ways to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy;

[0129] Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list.

[0130] For example, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party to the payment, and the sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list, including: using different binary values ​​to represent the sanction status identifier of each party to the payment to indicate whether the party is on the cross-border transaction sanction list.

[0131] For example, different payment scenario types correspond to different payment scenario type numbers. The two parties to the payment are the payer and the payee. The historical fraud score value of each party to the payment is the percentage of cross-border payment transactions that the party has participated in that were rated as fraudulent by the other party. The sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. It also includes: the cross-border transaction sanction list has an overflow limit on the maximum number of sanctions.

[0132] Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties.

[0133] In each training iteration of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the model, while the payment scenario type number, payment amount, payment-related content of the transaction, and multiple sales data are used as the input data of the deep cross-network fusion model to complete the training.

[0134] Specifically, numerical simulation mode can be used to test and simulate each training iteration of the gradient boosting decision tree and deep cross-network fusion model.

[0135] Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types;

[0136] For example, the numerical trends of the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model and the total number of existing transaction device types are the same, including: when there are 50 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1000; when there are 60 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1200; when there are 70 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1400; when there are 80 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1600; and so on.

[0137] like Figure 6 As shown, for example, N processors are given, where N is a natural number greater than or equal to 1.

[0138] Example 6

[0139] Figure 7 The diagram below illustrates a multi-dimensional encrypted transmission protection system for cross-border payments according to Embodiment 6 of the present invention. The system can be applied to a payment management server that performs cross-border payment management.

[0140] like Figure 7 As shown, the multi-dimensional encrypted transmission protection system for cross-border payments includes the following components:

[0141] The first capturing device is used to collect in real time the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score value of both parties, and the sanction status identifier of both parties for each cross-border payment transaction, so as to output each payment-related content of the cross-border payment transaction.

[0142] Specifically, the transaction devices of both parties in each cross-border payment transaction can be tablets, laptops, mobile phones, desktops, servers, and other transaction device types, with different transaction device types corresponding to different transaction device type numbers;

[0143] Specifically, the IP geographic information of both parties in each cross-border payment transaction can be the geographic name of the region to which the IP addresses currently used by both parties belong. For example, it can be the ASCII code value corresponding to the geographic name of the region to which the IP addresses currently used by both parties belong.

[0144] The second capturing device is used to collect in real time the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction for each cross-border payment transaction, so as to output multiple product sales data for that cross-border payment transaction.

[0145] Specifically, for each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to serve as multiple product sales data outputs for that cross-border payment transaction. The products purchased for each cross-border payment transaction can be of various types, such as electronic facilities, luxury goods, and office supplies.

[0146] The intelligent parsing device, connected to the first and second capturing devices, is used to intelligently parse the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction, and multiple commodity sales data using a gradient boosting decision tree and deep cross-network fusion model. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0147] Specifically, the intelligent parsing provides five specific values ​​for the five dimensions of the current cross-border payment transaction's unique dynamic multi-dimensional encryption strategy: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier.

[0148] The message generation device, connected to the intelligent parsing device, is used to encrypt, sign, and encapsulate the payment request of the current cross-border payment transaction according to the dynamic multi-dimensional encryption strategy to form a payment message, and transmits the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy concurrently to the recipient's transaction device through multiple channels.

[0149] Specifically, the payment request of the current cross-border payment transaction is encrypted, signed, and encapsulated to form a payment message based on a dynamic multi-dimensional encryption strategy, and the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy are concurrently transmitted to the recipient's transaction device through multiple channels. The channel isolation level is used to securely isolate the multiple channels. The higher the channel isolation level, the more stringent the security isolation of the multiple channels, that is, the higher the encryption degree of the corresponding dynamic multi-dimensional encryption strategy.

[0150] The decryption processing device, located on the recipient's transaction device and connected to the message generation device, is used to reconstruct the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy in order to complete the decryption operation of the payment message.

[0151] Among the five dimensions of dynamic multidimensional encryption strategy, including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multidimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multidimensional encryption strategy.

[0152] For example, the quantum key distribution identifier used to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy includes: using different binary values ​​to represent the quantum key distribution identifier in different ways to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy;

[0153] Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list.

[0154] For example, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party to the payment, and the sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list, including: using different binary values ​​to represent the sanction status identifier of each party to the payment to indicate whether the party is on the cross-border transaction sanction list.

[0155] For example, different payment scenario types correspond to different payment scenario type numbers. The two parties to the payment are the payer and the payee. The historical fraud score value of each party to the payment is the percentage of cross-border payment transactions that the party has participated in that were rated as fraudulent by the other party. The sanction status identifier of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. It also includes: the cross-border transaction sanction list has an overflow limit on the maximum number of sanctions.

[0156] Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties.

[0157] In each training iteration of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the model, while the payment scenario type number, payment amount, payment-related content of the transaction, and multiple sales data are used as the input data of the deep cross-network fusion model to complete the training.

[0158] Specifically, numerical simulation mode can be used to test and simulate each training iteration of the gradient boosting decision tree and deep cross-network fusion model.

[0159] Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types;

[0160] For example, the numerical trends of the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model and the total number of existing transaction device types are the same, including: when there are 50 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1000; when there are 60 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1200; when there are 70 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1400; when there are 80 types of existing transaction devices, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model is 1600, and so on.

[0161] Furthermore, the present invention may also reference the following technical contents to further demonstrate the outstanding substantial progress of the present invention:

[0162] For each cross-border payment transaction, the product type number of the purchased goods, the total number of historical sales, the average historical sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to serve as multiple product sales data outputs for that cross-border payment transaction. This includes a positive correlation between the number of each historical cross-border payment transaction and the total number of existing transaction device types.

[0163] Specifically, for each cross-border payment transaction, the real-time collection of the product type number, historical total sales, historical average sales amount, and transaction-related information of each historical cross-border payment transaction to serve as multiple product sales data outputs for that cross-border payment transaction also includes: each historical cross-border payment transaction being the most recent historical cross-border payment transaction used to purchase the product.

[0164] For example, for each cross-border payment transaction, the real-time collection of the product type number of the purchased goods, the total historical sales, the historical average sales amount, and transaction-related information of each historical cross-border payment transaction as multiple product sales data outputs for that cross-border payment transaction also includes: using a numerical conversion formula to represent the numerical conversion relationship between the number of each historical cross-border payment transaction and the total number of existing transaction device types.

[0165] Alternatively, a deep neural network model can be used to replace the gradient boosting decision tree and deep cross-network fusion model, wherein the deep neural network model is a deep neural network that has been trained multiple times.

[0166] Having described the embodiments of the invention and their advantages, it should be noted that various changes, substitutions, and modifications can be made without departing from the spirit and scope of the invention as defined in the appended claims. Furthermore, the use of terms such as "first," "second," etc., does not indicate any order or importance, but is used to distinguish elements from each other.

Claims

1. A multi-dimensional encrypted transmission protection method for cross-border payments, characterized in that, The method includes: For each cross-border payment transaction, the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties are collected in real time to output the payment-related content of the cross-border payment transaction. For each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to output multiple sets of product sales data for that cross-border payment transaction. The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier. The payment request for the current cross-border payment transaction is encrypted, signed, and encapsulated according to the dynamic multi-dimensional encryption strategy to form a payment message. The payment message and the policy digest of the dynamic multi-dimensional encryption strategy are transmitted concurrently to the recipient's transaction device through multiple channels. The payee's transaction device reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy to complete the decryption operation of the payment message; Among them, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multi-dimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multi-dimensional encryption strategy. Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties. Among them, the number of each historical cross-border payment transaction is positively correlated with the total number of existing transaction device types, and a numerical conversion formula is used to represent the numerical conversion relationship between the number of each historical cross-border payment transaction and the total number of existing transaction device types. Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types.

2. The multi-dimensional encrypted transmission protection method for cross-border payments as described in claim 1, characterized in that: In each training iteration of the gradient boosting decision tree and deep cross-network fusion model, the dynamic multi-dimensional encryption strategy of a known historical cross-border payment transaction is used as the output data of the model. The payment scenario type number, payment amount, payment-related content of the historical cross-border payment transaction, and multiple sets of product sales data are used as the input data of the deep cross-network fusion model to complete the training.

3. The multi-dimensional encrypted transmission protection method for cross-border payments as described in claim 2, characterized in that, After employing a gradient boosting decision tree and deep cross-network fusion model, based on the payment scenario type number, payment amount, various payment-related contents of the current cross-border payment transaction, and multiple commodity sales data, the method intelligently parses the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction. This dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. The method further includes: Receive the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction, and store the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction on the payer's transaction device. Among them, receiving the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction and storing the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction on the payer's transaction device includes: using a dynamic storage chip to complete the storage of the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction. The process of receiving the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction and storing the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction on the payer's transaction device also includes setting a storage period for the storage of the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction.

4. The multi-dimensional encrypted transmission protection method for cross-border payments as described in claim 2, characterized in that, Before collecting the transaction device type numbers of both parties, their IP address information, historical fraud scores, and sanction status of each party in real time for each cross-border payment transaction to output the associated payment information for that transaction, the method further includes: The more types of transaction devices there are, the more times the gradient boosting decision tree and deep cross-network fusion model needs to be trained.

5. The multi-dimensional encrypted transmission protection method for cross-border payments as described in claim 2, characterized in that, After employing a gradient boosting decision tree and deep cross-network fusion model, based on the payment scenario type number, payment amount, various payment-related contents of the current cross-border payment transaction, and multiple commodity sales data, the method intelligently parses the unique dynamic multi-dimensional encryption strategy corresponding to the current cross-border payment transaction. This dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. The method further includes: The dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction is wirelessly transmitted to a remote encryption management server; The wireless transmission of the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction to the remote encryption management server includes: the wireless transmission being encrypted transmission.

6. The multi-dimensional encrypted transmission protection method for cross-border payments as described in any one of claims 2-5, characterized in that: In a dynamic multi-dimensional encryption strategy that includes five dimensions such as encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier, quantum key distribution utilizes the properties of quantum mechanics to generate and share a random key for both parties in a cross-border payment transaction. Quantum key distribution utilizes quantum mechanical properties to generate and share a random key for both parties in a cross-border payment transaction. This includes: if a third party outside the two parties attempts to eavesdrop on the key, the cryptographic analysis operation used to eavesdrop will generate detectable anomalies. These anomalies are transmitted through quantum superposition or quantum entanglement states, allowing the two parties to detect the eavesdropping. When the number of third parties eavesdropping on the generated random key is less than or equal to a set threshold, the generated random key is shared with both parties in the cross-border payment transaction.

7. The multi-dimensional encrypted transmission protection method for cross-border payments as described in any one of claims 2-5, characterized in that: The numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model is the same as that of the total number of existing transaction device types, including: using a total number change curve to represent the numerical trend of the total number of existing transaction device types, and using a number change curve to represent the numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model. Among them, the numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model being the same as the total number of existing transaction device types also includes: the curvature values ​​at uniform intervals on the total number change curve are equal to the curvature values ​​at uniform intervals on the number of training sessions change curve. Among them, the numerical trend of the number of training sessions performed on the gradient boosting decision tree and deep cross-network fusion model being the same as the total number of existing transaction device types also includes: using programmable logic devices to simulate and model the total number change curve and the number of training sessions change curve.

8. A multi-dimensional encrypted transmission protection system for cross-border payments, characterized in that, The system includes a memory and multiple processors. The memory stores a computer program configured to be executed by the multiple processors to perform the following steps: For each cross-border payment transaction, the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score of both parties, and the sanction status of both parties are collected in real time to output the payment-related content of the cross-border payment transaction. For each cross-border payment transaction, the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction are collected in real time to output multiple sets of product sales data for that cross-border payment transaction. The gradient boosting decision tree and deep cross-network fusion model is adopted to intelligently analyze the dynamic multi-dimensional encryption strategy that uniquely corresponds to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction and multiple commodity sales data. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier. The payment request for the current cross-border payment transaction is encrypted, signed, and encapsulated according to the dynamic multi-dimensional encryption strategy to form a payment message. The payment message and the policy digest of the dynamic multi-dimensional encryption strategy are transmitted concurrently to the recipient's transaction device through multiple channels. The processor simulates the recipient's transaction device and reconstructs the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy to complete the decryption operation of the payment message; Among the five dimensions of dynamic multidimensional encryption strategy, including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multidimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multidimensional encryption strategy. Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties. Among them, the number of each historical cross-border payment transaction is positively correlated with the total number of existing transaction device types, and a numerical conversion formula is used to represent the numerical conversion relationship between the number of each historical cross-border payment transaction and the total number of existing transaction device types. Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types.

9. A multi-dimensional encrypted transmission protection system for cross-border payments, characterized in that, The system includes: The first capturing device is used to collect in real time the transaction device type number of both parties, the IP address and geographic information of both parties, the historical fraud score value of both parties, and the sanction status identifier of both parties for each cross-border payment transaction, so as to output each payment-related content of the cross-border payment transaction. The second capturing device is used to collect in real time the product type number, historical sales volume, historical average sales amount, and transaction-related information of each historical cross-border payment transaction for each cross-border payment transaction, so as to output multiple product sales data for that cross-border payment transaction. The intelligent parsing device, connected to the first and second capturing devices, is used to intelligently parse the dynamic multi-dimensional encryption strategy uniquely corresponding to the current cross-border payment transaction based on the payment scenario type number, payment amount, payment-related content of the current cross-border payment transaction, and multiple commodity sales data using a gradient boosting decision tree and deep cross-network fusion model. The dynamic multi-dimensional encryption strategy includes five dimensions: encryption algorithm type, key length, key update frequency, channel isolation level, and quantum key distribution identifier. The message generation device, connected to the intelligent parsing device, is used to encrypt, sign, and encapsulate the payment request of the current cross-border payment transaction according to the dynamic multi-dimensional encryption strategy to form a payment message, and transmits the payment message and its accompanying strategy digest of the dynamic multi-dimensional encryption strategy concurrently to the recipient's transaction device through multiple channels. The decryption processing device, located on the recipient's transaction device and connected to the message generation device, is used to reconstruct the dynamic multi-dimensional encryption strategy based on the strategy digest of the dynamic multi-dimensional encryption strategy in order to complete the decryption operation of the payment message. Among the five dimensions of dynamic multidimensional encryption strategy, including encryption algorithm type, key length, key update frequency, channel isolation level and quantum key distribution identifier, the higher the channel isolation level, the higher the encryption level of the corresponding dynamic multidimensional encryption strategy. The quantum key distribution identifier is used to indicate whether quantum key distribution is used in the corresponding dynamic multidimensional encryption strategy. Among them, different payment scenario types correspond to different payment scenario type numbers, the two parties to the payment are the payer and the payee, the historical fraud score value of each party to the payment is the percentage of cross-border payment transactions in which the party has participated in the past that were rated as fraudulent by the other party, and the sanction status mark of each party to the payment is used to indicate whether the party is on the cross-border transaction sanction list. Among them, the transaction-related information of each historical cross-border payment transaction for purchasing goods is each payment-related content of the historical cross-border payment transaction for purchasing the goods. Each payment-related content of each historical cross-border payment transaction includes the transaction device type number of the two parties in the historical cross-border payment transaction, the IP geographical information of the two parties, the historical fraud score value of the two parties, and the sanction status identifier of the two parties. Among them, the number of each historical cross-border payment transaction is positively correlated with the total number of existing transaction device types, and a numerical conversion formula is used to represent the numerical conversion relationship between the number of each historical cross-border payment transaction and the total number of existing transaction device types. Among them, the number of training iterations performed on the gradient boosting decision tree and deep cross-network fusion model showed the same numerical trend as the total number of existing transaction device types.

Images