Data interaction system, method and related device based on remote model implantation

Abstract

The application discloses a data interaction system and method based on remote model implantation and related devices, which can be applied to the field of network security or the field of finance. A first privacy calculation node deployed on a first institution side and a second privacy calculation node deployed on a second institution side interact with private data, i.e., a security model, through a first channel; a first server deployed on the first institution side interacts with non-private data with a second server deployed on the second institution side through a second channel; the first server sends a data request to the second server through the second channel, the second server determines target data information according to the data request and sends the target data information to the second privacy calculation node, the second privacy calculation node obtains an evaluation result through the security model and sends the evaluation result to the first server through the second server. The evaluation result does not include private data of an enterprise, so that the purpose of data interaction between different institutions is achieved without leaking private data.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of network security, and more particularly, to a data interaction system and method based on remote model implantation and related devices. BACKGROUND

[0002] The data stored in the databases on different institution sides can be different, for example, the database on the bank institution side stores transaction data of an enterprise, and the database on the tax institution side stores financial report data of the enterprise. Data interaction needs to be performed between different institution sides.

[0003] How to realize data interaction between different institutions while ensuring that the privacy of the enterprise is not disclosed is a technical problem that needs to be solved urgently. SUMMARY

[0004] Therefore, the present application provides a data interaction system and method based on remote model implantation and related devices.

[0005] To achieve the above object, the present application provides the following technical solutions.

[0006] According to a first aspect of the embodiments of the present disclosure, a data interaction system based on remote model implantation is provided, comprising:

[0007] a first privacy computing node deployed on a first institution side, configured to send a pre-constructed security model to a second privacy computing node deployed on a second institution side through a first channel;

[0008] In the process of transmitting the security model, the first channel is in a one-way data transmission state, which means that the first privacy computing node transmits to the second privacy computing node. In the process of training the security model, the first channel is in a two-way data transmission state. In the case of not transmitting the security model or not training the security model, the first channel is in a disconnected state. The security model is used to obtain an evaluation result of data information input into the security model, and the evaluation result is composed of multiple characters.

[0009] a first server deployed on the first institution side, configured to send a data request to a second server deployed on the second institution side through a second channel, the data request comprising a target enterprise identifier and a target data category to be obtained;

[0010] The second server is configured to find target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories, and data information, and input the target data information into the second privacy computing node.

[0011] The second privacy computing node is further configured to input the target data information into the security model stored by the second privacy computing node to obtain an evaluation result, and send the evaluation result to the second server.

[0012] The second server is further configured to send the evaluation result to the first server through the second channel.

[0013] According to a second aspect of the embodiments of the present disclosure, a data interaction method based on remote model implantation is provided, applied to a second server deployed on a second institution side, and the data interaction method based on remote model implantation comprises the following steps of:

[0014] receiving a data request from a first server deployed on a first institution side through a second channel, the data request comprising a target enterprise identifier and a target data category to be obtained;

[0015] finding target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories and data information;

[0016] inputting the target data information into a second privacy computing node deployed on the second institution side;

[0017] wherein the second privacy computing node stores a security model, the security model is sent to the second privacy computing node by a first privacy computing node deployed on the first institution side through a first channel, in the process of transmitting the security model, the first channel is in a one-way data transmission state, the one-way data transmission state refers to transmission from the first privacy computing node to the second privacy computing node, in the process of training the security model, the first channel is in a two-way data transmission state, in the case of non-transmission of the security model or non-training of the security model, the first channel is in a disconnected state, the security model is used to obtain an evaluation result of data information input into the security model, and the evaluation result is composed of multiple characters;

[0018] receiving an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model;

[0019] sending the evaluation result to the first server through the second channel.

[0020] According to a third aspect of the embodiments of the present disclosure, a data interaction device based on remote model implantation is provided, applied to a second server deployed on a second institution side, and the data interaction device based on remote model implantation comprises the following steps of:

[0021] The first receiving module is configured to receive a data request from a first server deployed on a first institution side through a second channel, the data request comprising a target enterprise identifier and a target data category to be obtained;

[0022] The searching module is configured to search for target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence between enterprise identifiers, data categories, and data information.

[0023] The input module is configured to input the target data information to a second privacy computing node deployed on the second institution side.

[0024] The second privacy computing node stores a security model, which is sent to the second privacy computing node by a first privacy computing node deployed on the first institution side through a first channel. During transmission of the security model, the first channel is in a one-way data transmission state, which means that data is transmitted from the first privacy computing node to the second privacy computing node. During training of the security model, the first channel is in a two-way data transmission state. In the case of non-transmission of the security model or non-training of the security model, the first channel is in a disconnected state. The security model is used to obtain an evaluation result of data information input to the security model, and the evaluation result is composed of multiple characters.

[0025] The second receiving module is configured to receive an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model.

[0026] The sending module is configured to send the evaluation result to the first server through the second channel.

[0027] According to a fourth aspect of the embodiments of the present disclosure, a server is provided, which comprises a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps of the method of the second aspect.

[0028] According to a fifth aspect of the embodiments of the present disclosure, a computer readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the method of the second aspect are implemented.

[0029] According to a sixth aspect of the embodiments of the present disclosure, a computer program product is provided, which comprises a computer program. When the computer program is executed by a processor, the steps of the method of the second aspect are implemented.

[0030] Via the technical solution, the application provides a data interaction system based on remote model implantation. The first privacy calculation node deployed on the first institution side and the second privacy calculation node deployed on the second institution side interact with the private data through the first channel. In the process of transmitting the security model, the second privacy calculation node cannot transmit data to the first privacy calculation node, thereby avoiding the case that the enterprise data stored in the database of the second institution side is transmitted to the first institution side. If the first institution side needs the private data of the enterprise stored in the database of the second institution side, at this time, the first channel is in a disconnected state, the first server deployed on the first institution side sends a data request to the second server deployed on the second institution side through the second channel, the data request includes a target enterprise identifier and a target data category to be obtained; the second server searches for target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories and data information; the target data information is input into the second privacy calculation node; the second privacy calculation node inputs the target data information into the security model stored in the second privacy calculation node to obtain an evaluation result; the evaluation result is sent to the second server; and the second server sends the evaluation result to the first server through the second channel. Since the evaluation result is a label or an evaluation level obtained from the target data information by the security model, the evaluation result does not include the private data of the enterprise, thereby achieving the purpose of data interaction between different institutions without leaking the private data. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0032] Figure 1 is a schematic diagram of a hardware architecture related to the present application according to an exemplary embodiment;

[0033] Figure 2 is a flowchart of a data interaction system based on remote model implantation according to an exemplary embodiment;

[0034] Figure 3 is a flowchart of a data interaction method based on remote model implantation according to an exemplary embodiment;

[0035] Figure 4 is a block diagram of a data interaction device based on remote model implantation according to an exemplary embodiment;

[0036] Figure 5is a block diagram of an apparatus for a server according to an exemplary embodiment. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0038] It should be noted that the enterprise information (including but not limited to enterprise device information, enterprise personal information, etc.) and data (including but not limited to data for analysis, stored data, transaction data, etc.) involved in the present application are all information and data authorized by the enterprise or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0039] The embodiments of the present application provide a data interaction system and method based on remote model implantation and related devices. Before the technical solutions provided by the embodiments of the present application, the application scenarios and hardware architectures involved in the present application are described.

[0040] First, the application scenarios involved in the present application are described.

[0041] The databases of different institution sides store different data, and data interaction is needed between different institution sides, for example, the bank institution side needs to obtain the financial data of enterprises stored in the database of the tax institution side. The data interaction needs to solve two problems, the first problem is how to use data to realize the value of industrial development and build mutual trust between different institution sides, and the second problem is to ensure the privacy of data and protect the data security requirements of different institution sides. The above two problems are urgent problems to be solved in data development and sharing.

[0042] Secondly, the hardware architecture involved in the present application is described.

[0043] Figure 1 is a schematic diagram of the hardware architecture involved in the present application according to an exemplary embodiment, which includes a first privacy computing node 11 deployed at a first institution side, a first server 12 deployed at the first institution side, a second privacy computing node 13 deployed at a second institution side, and a second server 14 deployed at the second institution side.

[0044] Exemplarily, the hardware architecture further includes a first switch 15 deployed at the first institution side, a first router 16 deployed at the first institution side, a second switch 17 deployed at the second institution side, and a second router 18 deployed at the second institution side.

[0045] Exemplarily, the first privacy computing node or the second privacy computing node can be any electronic product that can interact with a user through one or more methods such as a keyboard, touchpad, touch screen, remote control, voice interaction or handwriting device, such as a mobile phone, tablet computer, PDA, personal computer, wearable device, etc.

[0046] Exemplarily, the first privacy computing node or the second privacy computing node can be a server, a server cluster consisting of multiple servers, or a cloud computing service center.

[0047] Exemplarily, the first privacy computing node or the second privacy computing node can be a hot standby dual-machine.

[0048] Illustratively, the first server or the second server may be a single server, or a server cluster consisting of multiple servers, or a cloud computing service center.

[0049] Exemplarily, the first server or the second server may be in hot standby mode.

[0050] Illustratively, the first institution side may be a bank side of a banking institution, and the second institution side may be a relevant government department, such as a tax agency side.

[0051] Exemplarily, when the first institution side and the second institution side transmit private data, they transmit through the first channel 19. When the first institution side and the second institution side transmit non-private data, they transmit through the second channel.

[0052] like Figure 1 As shown, the second channel is composed of a first server 12 , a first switch 15 , a first router 16 , a wireless network, a second router 18 , a second switch 17 and a second server 14 .

[0053] Exemplarily, the first channel may be a wired connection channel or a wireless connection channel.

[0054] Exemplarily, the first channel is open only during the transmission of private data and is disconnected at other times. Exemplarily, the first channel can be manually opened and disconnected offline. Exemplarily, the first channel can be opened and disconnected online after authorization is passed.

[0055] For example, the communication channel between the first privacy computing node 11 and the first server 12 is open when transmitting private data, and is disconnected at other times; the communication channel between the second privacy computing node 13 and the second server 14 is open when transmitting private data, and is disconnected at other times.

[0056] In an optional implementation, the database of the second institution side stores the privacy data of the enterprise required by the first institution side, and the second privacy computing node stores the security model obtained by the first institution side. The first privacy computing node and the second privacy computing node are both deployed by the first institution (corresponding to the first institution side), and the second institution side has no permission to change or view the security model stored in the second privacy computing node, so as to achieve the purpose that the security model stored in the second privacy computing node is not changed by the second institution side. Illustratively, the second institution side has the permission to view the log of the second privacy computing node. Illustratively, the second institution side has the permission to determine the data transmission direction and data transmission frequency between the first privacy computing node and the second privacy computing node.

[0057] In summary, in the embodiment of the present application, the privacy data is transmitted between the first privacy computing node 11 and the second privacy computing node 13 through the first channel 19. The first channel 19 is only turned on in the case of transmitting privacy data, and is in a disconnected state at other times. The non-private data is transmitted between the first server 12 and the second server 14 through the second channel.

[0058] Illustratively, during the transmission of the privacy data between the first privacy computing node 11 and the second privacy computing node 13 through the first channel 19, the communication channel between the first server 12 and the first privacy computing node 11 is disconnected, and the communication channel between the second server 14 and the second privacy computing node 13 is disconnected.

[0059] It should be understood by those skilled in the art that the above device is only an example, and other existing or future devices that can be applicable to the present disclosure should also be included in the protection scope of the present disclosure, and are hereby incorporated by reference.

[0060] The data interaction system based on remote model implantation provided by the embodiment of the present application will be described below in combination with the hardware architecture described above.

[0061] Figure 2 is a flowchart of a data interaction system based on remote model implantation according to an exemplary embodiment, as shown in Figure 2 includes the following steps S21 to S25.

[0062] Step S21: The first privacy computing node deployed at the first institution side sends the pre-constructed security model to the second privacy computing node deployed at the second institution side through the first channel.

[0063] In which, during the transmission of the security model, the first channel is in a unidirectional data transmission state, and the unidirectional data transmission state refers to transmission from the first privacy computing node to the second privacy computing node; during the training of the security model, the first channel is in a bidirectional data transmission state; when the security model is not transmitted or trained, the first channel is in a disconnected state; the security model is used to obtain an evaluation result of the data information input into the security model, and the evaluation result is composed of multiple characters.

[0064] It can be understood that during the execution of step S21, the first privacy computing node can transmit data to the second privacy computing node, but the second privacy computing node cannot transmit data to the first privacy computing node, thereby avoiding the situation where the second privacy computing node transmits the private data stored on the second institution side to the first institution.

[0065] Exemplarily, the process of transmitting the security model and training the security model is the process of transmitting private data, during which the first channel is in an on state and is in an off state at other times.

[0066] For example, a security model can be trained using federated learning. During the training of the security model, the first and second privacy-preserving computing nodes need to exchange data, so the first channel is in a bidirectional data transmission state. That is, the first privacy-preserving computing node can transmit data to the second privacy-preserving computing node through the first channel, and the second privacy-preserving computing node can transmit data to the first privacy-preserving computing node through the first channel.

[0067] For example, the number of security models may be one or more. Different security models have different functions, and the sample data for training security models with different functions may be different.

[0068] Exemplarily, the security model may be a machine learning model or a generative large language model.

[0069] For example, the second institution (corresponding to the second institution side) can be a government department. The second institution may collect production and operation data of enterprises during administrative management. This data is the private data of the enterprises. Production and operation data includes, but is not limited to, financial reports, water and electricity data, and other private data that is not suitable for public disclosure in the financial market.

[0070] In order to prevent the company's private data from being obtained by the first institution, if the first institution needs to use this private data, it can train a security model. The security model can extract the labels or ratings of this private data. The first institution only needs to obtain the labels or ratings without obtaining the company's private data.

[0071] In an optional implementation, the process in which the security model extracts the label or score level of the privacy data is actually a process of grading or evaluating the privacy data, for example, the security model can be a financial report use model, and the financial report use model can obtain the evaluation result of the enterprise based on the financial report data of the enterprise. Assuming that the evaluation result is an evaluation level, for example, the evaluation level includes a first level, a second level, and a third level. Illustratively, the enterprise needs to recognize the evaluation level output by the financial report use model.

[0072] It can be understood that, in order to avoid the second institution inferring the evaluation method of the first institution based on the evaluation result output by the security model, that is, the way in which the security model obtains the label or score level of the privacy data, the evaluation result output by the security model can be represented by a specified character, and the second institution does not know the meaning of the character, so that the second institution can avoid inferring the evaluation method of the first institution based on the evaluation result output by the security model.

[0073] Illustratively, the evaluation result output by the security model is composed of multiple characters, for example, characters A, B, C, D, and E.

[0074] Illustratively, the first institution side is a bank institution side, and the second institution side is a government-related institution side. Thus, the security requirements of the three-party data of banks, governments, and enterprises are constructed. The second institution side can belong to the government cloud side.

[0075] Illustratively, the first privacy computing node and the second privacy computing node use a privacy computing mode. Privacy compute or privacy computing refers to a collection of technologies that realize data analysis and calculation while protecting data from being leaked to the outside, so as to achieve the purpose of “available and invisible” of data; under the premise of fully protecting data and privacy security, the transformation and release of data value are realized.

[0076] Step S22: The first server deployed on the first institution side sends a data request to the second server deployed on the second institution side through a second channel, and the data request includes a target enterprise identifier and a target data category to be obtained.

[0077] Illustratively, the target enterprise identifier can be the name of the enterprise or the unified social credit code of the enterprise.

[0078] Illustratively, the target data category refers to the category of privacy data that needs to be evaluated, such as a financial data category, a tax data category, and a loan data category.

[0079] Step S23: The second server finds the target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories, and data information; and inputs the target data information into the second privacy computing node.

[0080] For example, the database deployed on the second institution side stores the correspondence relationship among enterprise identifiers, data categories, and data information.

[0081] Step S24: The second privacy computing node inputs the target data information into the security model stored in the second privacy computing node to obtain an evaluation result; and sends the evaluation result to the second server.

[0082] Step S25: The second server sends the evaluation result to the first server through the second channel.

[0083] For example, if the second channel is in an on state, the first channel is in an off state, so the first privacy computing node 11 and the second privacy computing node 13 cannot interact with each other in the process of executing steps S23 to S25, and the second server sends the evaluation result to the first server 12, thereby avoiding the situation of sending the private data of the enterprise stored in the database of the second institution side to the first server 12.

[0084] The embodiment of the present application provides a data interaction system based on remote model implantation. The first privacy computing node deployed on the first institution side and the second privacy computing node deployed on the second institution side interact with each other through a first channel to transmit private data. In the process of transmitting the security model, the second privacy computing node cannot transmit data to the first privacy computing node, thereby avoiding the situation of transmitting the private data of the enterprise stored in the database of the second institution side to the first institution side. If the first institution side needs the private data of the enterprise stored in the database of the second institution side, at this time, the first channel is in an off state, the first server deployed on the first institution side sends a data request to the second server deployed on the second institution side through a second channel, the data request includes a target enterprise identifier and a target data category to be obtained; the second server finds the target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories, and data information; and inputs the target data information into the second privacy computing node; the second privacy computing node inputs the target data information into the security model stored in the second privacy computing node to obtain an evaluation result; and sends the evaluation result to the second server; and the second server sends the evaluation result to the first server through the second channel. Since the evaluation result is a label or an evaluation grade obtained by the security model from the target data information, the evaluation result does not include the private data of the enterprise, thereby achieving the purpose of realizing data interaction between different institutions without leaking the private data.

[0085] In order to ensure that the evaluation result output by the security model does not include privacy data, further protecting the privacy data of the enterprise, the embodiment of the application further provides the following method, which comprises the following steps A1 to A2.

[0086] Step A1: If the length of the evaluation result is less than or equal to a preset length, and all characters in the evaluation result belong to a preset character set, it is determined that the evaluation result meets the requirements.

[0087] Step A2: If the length of the evaluation result is greater than the preset length, or at least one character in the evaluation result does not belong to the preset character set, it is determined that the evaluation result does not meet the requirements.

[0088] It can be understood that the first institution side and the second institution side can agree on the length of the evaluation result in advance, for example, 10 bytes, and can agree on the content of each byte in advance, for example, each byte can only be represented by A, B, C, D, and E, so the preset length can be 10, and the preset character set includes A, B, C, D, and E.

[0089] For example, the evaluation result is a part of the message representing the label or evaluation level in the feedback from the second server to the first server.

[0090] It can be understood that if the evaluation result meets the requirements, step S25 can be performed. Otherwise, step S25 is not performed.

[0091] It can be understood that the second server will send the target data information to the second privacy computing node, and if the second privacy computing node stores the target data information, the second privacy computing node may send the target data information to the first privacy computing node in the subsequent process of training the security model, resulting in leakage of the privacy data of the enterprise. In order to avoid this situation, the embodiment of the application provides the following method, which comprises the following step B1.

[0092] Step B1: The electronic device deployed on the second institution side acquires the log of the second privacy computing node, the log comprising write information and read information of the second privacy computing node for a hard disk; and determines whether the second privacy computing node stores the target data information based on the log.

[0093] The hard disk is provided with a write protection function.

[0094] For example, the electronic device is any kind of electronic product that can interact with a user through one or more of a keyboard, a touchpad, a touch screen, a remote control, voice interaction, or a handwriting device, such as a mobile phone, a tablet computer, a palm computer, a personal computer, a wearable device, etc.

[0095] Exemplarily, if the second privacy computing node wants to store the target data information, the second privacy computing node needs to write the target data information into the hard disk. Since the hard disk is provided with the write protection function, the target data information cannot be written into the hard disk, so that the second privacy computing node cannot store the target data information.

[0096] Exemplarily, since the hard disk has the write protection function, the system administrator needs to authorize before data can be written into the hard disk.

[0097] Exemplarily, the write information includes the time when the second privacy computing node writes data into the hard disk, the data size, and the write address. Exemplarily, the read information includes the read time when the second privacy computing node reads data from the hard disk, and the address where the data is located. Thus, whether the second privacy computing node stores the privacy data of the enterprise can be obtained based on the write information and the read information.

[0098] Figure 3 is a flowchart of a data interaction method based on remote model implantation according to an exemplary embodiment. The method can be applied to a second server deployed on a second institution side. The method includes the following steps S31 to S35.

[0099] Step S31: receiving a data request from a first server deployed on a first institution side through a second channel, the data request including a target enterprise identifier and a target data category to be obtained.

[0100] Step S32: searching for target data information corresponding to the target enterprise identifier and the target data category from a pre-stored correspondence relationship among enterprise identifiers, data categories, and data information.

[0101] Step S33: inputting the target data information into a second privacy computing node deployed on the second institution side.

[0102] The second privacy computing node stores a security model. The security model is sent to the second privacy computing node by a first privacy computing node deployed on the first institution side through a first channel. During transmission of the security model, the first channel is in a one-way data transmission state, which means transmission from the first privacy computing node to the second privacy computing node. During training of the security model, the first channel is in a two-way data transmission state. In the case of non-transmission of the security model or non-training of the security model, the first channel is in a disconnected state. The security model is used to obtain an evaluation result of data information input into the security model, and the evaluation result is composed of multiple characters.

[0103] Step S35: receiving an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model.

[0104] Step S36: sending the evaluation result to the first server through the second channel.

[0105] The method is described in detail in the embodiments of the present disclosure. The method of the present disclosure can be implemented in various forms of devices. Therefore, the present disclosure also discloses a device. The following specific embodiments are given to explain in detail.

[0106] Figure 4 is a data interaction device block diagram based on remote model implantation according to an exemplary embodiment, applied to a second server deployed on a second institution side. Referring to Figure 4 , the device comprises a first receiving module 41, a searching module 42, an input module 43, a second receiving module 44 and a sending module 45, wherein:

[0107] The first receiving module 41 is configured to receive a data request from a first server deployed on a first institution side through a second channel, wherein the data request comprises a target enterprise identifier and a target data category to be obtained.

[0108] The searching module 42 is configured to search for target data information corresponding to the target enterprise identifier and the target data category from a pre-stored corresponding relationship among enterprise identifiers, data categories and data information.

[0109] The input module 43 is configured to input the target data information to a second privacy computing node deployed on the second institution side.

[0110] The second privacy computing node stores a security model, wherein the security model is sent to the second privacy computing node by a first privacy computing node deployed on the first institution side through a first channel. During transmission of the security model, the first channel is in a one-way data transmission state, which means that the first privacy computing node transmits to the second privacy computing node. During training of the security model, the first channel is in a two-way data transmission state. In the case of non-transmission of the security model or non-training of the security model, the first channel is in a disconnected state. The security model is used to obtain an evaluation result of data information input into the security model, and the evaluation result is composed of multiple characters.

[0111] The second receiving module 44 is configured to receive an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model.

[0112] The sending module 45 is configured to send the evaluation result to the first server through the second channel.

[0113] With regard to the apparatus in the above-described embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments of the method, and thus will not be described in detail here.

[0114] Figure 5 is a block diagram of an apparatus for a server according to an exemplary embodiment. Exemplarily, the server can be a second server.

[0115] The server includes, but is not limited to, a processor 51, a memory 52, a network interface 53, an I / O controller 54, and a communication bus 55.

[0116] It should be noted that those skilled in the art can understand that the structure of the server shown in the above embodiments does not constitute a limitation on the server, and the server can include more or fewer components than those shown in the above embodiments, or combine some components, or different component arrangements. Figure 5 The structure of the server shown in the above embodiments does not constitute a limitation on the server, and the server can include more or fewer components than those shown in the above embodiments, or combine some components, or different component arrangements. Figure 5 The structure of the server shown in the above embodiments does not constitute a limitation on the server, and the server can include more or fewer components than those shown in the above embodiments, or combine some components, or different component arrangements.

[0117] The following will be specifically introduced with reference to the Figure 5 The following will be specifically introduced with reference to the

[0118] The processor 51 is the control center of the server, and connects each part of the server through various interfaces and lines, and performs various functions of the server and processes data by running or executing software programs and / or modules stored in the memory 52 and calling data stored in the memory 52, thereby monitoring the server as a whole. The processor 51 can include one or more processing units; exemplarily, the processor 51 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 51.

[0119] The processor 51 can be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present application, etc.

[0120] The memory 52 can include a memory, such as a high-speed Random-Access Memory (RAM) 521 and a Read-Only Memory (ROM) 522, and can also include a mass storage device 523, such as at least one disk memory or the like. Of course, the server can also include other hardware required by the business.

[0121] The memory 52 described above is configured to store instructions executable by the processor 51 described above. The processor 51 described above is configured to execute the data interaction method based on remote model implantation.

[0122] A wired or wireless network interface 53 is configured to connect the server to a network.

[0123] The processor 51, the memory 52, the network interface 53, and the I / O controller 54 can be connected to each other through a communication bus 55, which can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0124] In an exemplary embodiment, the server can be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements, for executing the data interaction method based on remote model implantation described above.

[0125] In an exemplary embodiment, the present disclosure provides a storage medium, such as the memory 52 including instructions, which can be executed by the processor 51 of the server to complete the method described above. Alternatively, the storage medium can be a non-transitory computer-readable storage medium, such as a ROM, a Random-Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0126] In an exemplary embodiment, a computer-readable storage medium is also provided, which can be directly loaded into an internal memory of a computer, such as the memory 52 described above, and contains software code. The computer program can be loaded and executed by a computer to implement the data interaction method based on remote model implantation described above.

[0127] In an example embodiment, a computer program product is also provided, which can be directly loaded into the internal memory of a computer, such as the memory included in the server, and contains software codes. The computer program can be loaded and executed by a computer to implement the above-mentioned data interaction method based on remote model implantation.

[0128] It should be noted that the data interaction system, method and related device based on remote model implantation provided by the present application can be used in the field of network security or the field of finance. The above is only an example and does not limit the application field of the data interaction system, method and related device based on remote model implantation provided by the present application.

[0129] It should be noted that the features described in each of the embodiments in the present specification can be replaced or combined with each other. For device or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant part can be referred to the part of the description of the method embodiments.

[0130] It should also be noted that in this document, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or device including the element.

[0131] The steps of the method or algorithm described in connection with the embodiments disclosed herein can be directly implemented by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0132] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Modifications of these embodiments will occur to persons of skill in the art, and that the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data interaction system based on remote model implantation, characterized in that: include: A first privacy computing node deployed on the first institution side is used to send the pre-built security model to a second privacy computing node deployed on the second institution side through the first channel; During the transmission of the security model, the first channel is in a unidirectional data transmission state, where the unidirectional data transmission state refers to transmission from the first privacy computing node to the second privacy computing node. During the training of the security model, the first channel is in a bidirectional data transmission state. When the security model is not being transmitted or trained, the first channel is disconnected. The security model is used to obtain an evaluation result of the data information input into the security model, where the evaluation result is composed of multiple characters. A first server deployed at the first organization side is configured to send a data request to a second server deployed at the second organization side through a second channel, wherein the data request includes a target enterprise identifier and a target data category to be obtained; The second server is configured to search for target data information corresponding to the target enterprise identifier and the target data category from pre-stored correspondences between enterprise identifiers, data categories, and data information; and input the target data information into the second privacy computing node; The second privacy computing node is further configured to input the target data information into the security model stored in the node to obtain an evaluation result; and send the evaluation result to the second server; The second server is further configured to send the evaluation result to the first server through the second channel.

2. The data interaction system based on remote model implantation according to claim 1, characterized in that: Before executing the operation of sending the evaluation result to the first server through the second channel, the second server is further configured to: If the length of the evaluation result is less than or equal to the preset length, and the characters in the evaluation result all belong to the preset character set, it is determined that the evaluation result meets the requirements; If the length of the evaluation result is longer than the preset length, or at least one character in the evaluation result does not belong to the preset character set, it is determined that the evaluation result does not meet the requirements.

3. The data interaction system based on remote model implantation according to claim 1 or 2, characterized in that: Also includes: An electronic device deployed on the second institution side is used to obtain a log of the second privacy-preserving computing node, wherein the log includes information written to and read from the hard disk by the second privacy-preserving computing node; and determine whether the second privacy-preserving computing node stores the target data information based on the log; Wherein, the hard disk is provided with a write protection function.

4. The data interaction system based on remote model implantation according to claim 1, characterized in that: The second privacy computing node and the first privacy computing node both belong to the first organization.

5. The data interaction system based on remote model implantation according to claim 1 is characterized in that: The second channel is constructed by a first switch and a first router deployed on the first institution side, and a second switch and a second router deployed on the second institution side, as well as a wireless communication network.

6. A data interaction method based on remote model implantation, characterized in that: Applied to a second server deployed on the second institution side, the data interaction method based on remote model implantation includes: receiving, through a second channel, a data request from a first server deployed on the first organization side, the data request including a target enterprise identifier and a target data category to be obtained; Searching for target data information corresponding to the target enterprise identifier and the target data category from pre-stored correspondences among enterprise identifiers, data categories, and data information; Inputting the target data information into a second privacy computing node deployed on the second institution side; The second privacy computing node stores a security model; the security model is sent to the second privacy computing node by the first privacy computing node deployed on the first institution side through the first channel; during the transmission of the security model, the first channel is in a unidirectional data transmission state, and the unidirectional data transmission state refers to the transmission from the first privacy computing node to the second privacy computing node; during the training of the security model, the first channel is in a bidirectional data transmission state; when the security model is not being transmitted or trained, the first channel is in a disconnected state; the security model is used to obtain an evaluation result of the data information input to the security model, and the evaluation result is composed of multiple characters; Receiving an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model; The evaluation result is sent to the first server through the second channel.

7. A data interaction device based on remote model implantation, characterized in that: Applied to a second server deployed on the second institution side, the data interaction device based on remote model implantation includes: A first receiving module is configured to receive a data request from a first server deployed on the first organization side through a second channel, wherein the data request includes a target enterprise identifier and a target data category to be obtained; A search module, configured to search for target data information corresponding to the target enterprise identifier and the target data category from pre-stored correspondences among enterprise identifiers, data categories, and data information; An input module, configured to input the target data information into a second privacy computing node deployed on the second institution side; The second privacy computing node stores a security model; the security model is sent to the second privacy computing node by the first privacy computing node deployed on the first institution side through the first channel; during the transmission of the security model, the first channel is in a unidirectional data transmission state, and the unidirectional data transmission state refers to the transmission from the first privacy computing node to the second privacy computing node; during the training of the security model, the first channel is in a bidirectional data transmission state; when the security model is not being transmitted or trained, the first channel is in a disconnected state; the security model is used to obtain an evaluation result of the data information input to the security model, and the evaluation result is composed of multiple characters; A second receiving module is configured to receive an evaluation result corresponding to the target data information obtained by the second privacy computing node through the security model; A sending module is used to send the evaluation result to the first server through the second channel.

8. A server comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to claim 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 6 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to claim 6 are implemented.

Images