Privacy data processing method, device, equipment, storage medium and program product

Abstract

The present invention discloses a privacy data processing method, device, equipment, storage medium and program product. The method is applied to a data demand node in a member node of a blockchain. The blockchain includes a central node and multiple member nodes. The method includes: sending data demand information to the central node so that the central node determines a data providing node in a member node based on the data demand information, and sending the data demand information to the data providing node; in response to a transaction order sent by the data providing node, generating task configuration information based on the transaction order, and determining participating nodes; allocating data processing tasks to each participating node according to the task configuration information, the data processing tasks including allocation parameters corresponding to each participating node; managing data asset permissions based on the processing results of the data processing tasks fed back by each participating node, thereby ensuring efficient data management while effectively alleviating data processing pressure and realizing full life cycle management of data assets.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a privacy data processing method, device, equipment, storage medium and program product. Background Art

[0002] Data elements are an intangible economic resource that is closely related to digital technology and big data. With the widespread application of computers and Internet technologies, the activities of each economic entity in the market will generate a variety of initial data. Each user will rely on its own conditions and existing computing platforms to process this data through collection, storage, cleaning and other means to generate structured data products for use by various enterprises or users. However, due to factors such as the complex identities and business relationships between users, the wide variety of data, and the huge amount of data, it is currently difficult to achieve the full life cycle management of data assets, and data processing is difficult, and it is impossible to effectively alleviate the pressure of data processing. Summary of the invention

[0003] The main purpose of the present invention is to provide a privacy data processing method, device, equipment, storage medium and program product, aiming to solve the technical problems that the existing technology is difficult to achieve full life cycle management of data assets, and data processing is difficult and cannot effectively alleviate data processing pressure.

[0004] To achieve the above object, the present invention provides a privacy data processing method, which is applied to a data demand node with data demand among member nodes of a blockchain, wherein the blockchain includes a central node and multiple member nodes, and the privacy data processing method includes the following steps:

[0005] Sending data demand information to the central node, so that the central node determines a data providing node among the member nodes based on the data demand information, and sending the data demand information to the data providing node;

[0006] In response to the transaction order sent by the data providing node, generating task configuration information based on the transaction order, and determining participating nodes;

[0007] Allocate a data processing task to each participating node according to the task configuration information, wherein the data processing task includes allocation parameters corresponding to each participating node;

[0008] The data asset authority is managed based on the processing result of the data processing task fed back by each participating node according to the allocation parameter.

[0009] Optionally, in response to the transaction order sent by the data providing node, generating task configuration information based on the transaction order, and determining participating nodes, includes:

[0010] In response to a transaction order sent by the data providing node, obtaining providing node information according to the transaction order;

[0011] Performing an interconnection test on the data providing node according to the providing node information;

[0012] In response to the interconnection test being successful, generating a data processing work order based on the transaction order;

[0013] Sending the data processing work order to the data providing node so that the data providing node authorizes the data processing work order;

[0014] In response to the data providing node agreeing to the authorization, task configuration information is generated based on the data processing work order, and participating nodes are determined.

[0015] Optionally, the managing the data asset authority based on the processing result of the data processing task fed back by each participating node according to the allocation parameter includes:

[0016] Monitoring the processing status of the data processing task fed back by each participating node;

[0017] Based on the processing status, obtain the operation parameters fed back by each participating node;

[0018] Aggregate the job parameters fed back by each participating node to obtain the global parameters;

[0019] Encrypting the global parameter and sending the encrypted global parameter to the data providing node;

[0020] When the processing state is a processing completion state, determining a task processing weight of each participating node based on a processing result of the data processing task fed back by each participating node;

[0021] Allocating data usage rights to each participating node based on the task processing weight;

[0022] Manage data asset permissions based on data usage rights allocation results.

[0023] In addition, to achieve the above-mentioned purpose, the present invention also proposes a privacy data processing method, which is applied to a participating node participating in data processing among the member nodes of the blockchain, and the privacy data processing method includes:

[0024] In response to a data processing task sent by a data demand node, obtaining allocation parameters based on the data processing task;

[0025] Perform data processing based on the allocation parameters to obtain task processing parameters;

[0026] Encrypting the task processing parameters to obtain operation parameters;

[0027] The job parameters are sent to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node.

[0028] Optionally, encrypting the task processing parameters to obtain the operation parameters includes:

[0029] Generate a local public key and obtain the public key of each participating node in the member node of the blockchain;

[0030] Aggregate the local public key and the public keys of the participating parties of each participating node to obtain an aggregated public key;

[0031] The task processing parameters are encrypted according to the aggregate public key to obtain operation parameters.

[0032] Optionally, sending the job parameters to the data demand node so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node includes:

[0033] Sending the operation parameters to the data demand node, so that the data demand node aggregates the operation parameters sent by each participating node into a global parameter;

[0034] Obtaining global parameters sent by the data demand node;

[0035] Obtaining a decrypted segment based on the global parameter;

[0036] Adding noise information to the decrypted fragment based on a preset noise space to obtain a target decrypted fragment;

[0037] The target decrypted slice is sent to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the target decrypted slice sent by each participating node.

[0038] In addition, to achieve the above-mentioned purpose, the present invention also proposes a privacy data processing device, which is used to implement the steps of the privacy data processing method as described above.

[0039] In addition, to achieve the above-mentioned purpose, the present application also proposes a data processing device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the privacy data processing method as described above.

[0040] In addition, to achieve the above objectives, the present application also proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the privacy data processing method described above are implemented.

[0041] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the privacy data processing method described above.

[0042] The present invention sends data demand information to the central node so that the central node determines the data providing node among the member nodes based on the data demand information, and sends the data demand information to the data providing node; in response to the transaction order sent by the data providing node, generates task configuration information based on the transaction order, and determines the participating nodes; allocates data processing tasks to each participating node according to the task configuration information, and the data processing tasks include allocation parameters corresponding to each participating node; manages data asset permissions based on the processing results of the data processing tasks fed back by each participating node according to the allocation parameters; because the present invention allocates data processing tasks to each member node in the blockchain through the blockchain central node, the management of data processing tasks is realized by the central node, and the member nodes undertake their own data processing tasks according to the supply and demand parties, thereby ensuring efficient data management while effectively alleviating data processing pressure, and managing data asset permissions according to the processing results of each participating node, thereby realizing full life cycle management of data assets. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 It is a structural diagram of a data processing device in a hardware operating environment involved in an embodiment of the present invention;

[0046] Figure 2 This is a flowchart of a first embodiment of a privacy data processing method of the present invention;

[0047] Figure 3 This is a flow chart of a second embodiment of the privacy data processing method of the present invention;

[0048] Figure 4 This is a schematic diagram of the encryption calculation process in the second embodiment of the privacy data processing method of the present invention;

[0049] Figure 5 It is a structural block diagram of the first embodiment of the privacy data processing device of the present invention;

[0050] Figure 6 It is a structural block diagram of a central computing device in the first embodiment of the privacy data processing device of the present invention;

[0051] Figure 7 It is a structural block diagram of a member computing device in the first embodiment of the privacy data processing device of the present invention;

[0052] Figure 8 This is a structural block diagram of the universal privacy computing base module in the first embodiment of the privacy data processing device of the present invention.

[0053] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0055] Reference Figure 1 , Figure 1 The present invention is a schematic diagram of the data processing device structure of the hardware operating environment involved in the embodiment of the present invention.

[0056] like Figure 1 As shown, the data processing device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0057] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the data processing device, and may include more or less components than shown in the figure, or combine some components, or arrange the components differently.

[0058] like Figure 1 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a privacy data processing program.

[0059] exist Figure 1 In the data processing device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the data processing device of the present invention can be set in the data processing device, and the data processing device calls the privacy data processing program stored in the memory 1005 through the processor 1001, and executes the privacy data processing method provided by the embodiment of the present invention.

[0060] The embodiment of the present invention provides a privacy data processing method, referring to Figure 2 , Figure 2 Schematic diagram of the process of the first embodiment of the privacy data processing method of the present invention.

[0061] In this embodiment, the privacy data processing method is applied to a data demand node with data demand among the member nodes of a blockchain, wherein the blockchain includes a central node and multiple member nodes, and the privacy data processing method includes the following steps:

[0062] Step S10: Sending data demand information to the central node, so that the central node determines a data providing node among the member nodes based on the data demand information, and sends the data demand information to the data providing node.

[0063] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, etc. The following takes a data processing device as an example to illustrate this embodiment and the following embodiments.

[0064] It should be noted that the data demand information may be a data computing task initiated by a data demander in a blockchain member node. Blockchain member nodes include data demand nodes and data provider nodes. The data demand node may be a data demander, and the data provider node may be a data provider.

[0065] It is understandable that the blockchain central node can assign roles based on the order requirements of data transactions sent by member nodes.

[0066] It should be noted that the data demand node as the demand side can initiate computing tasks through the privacy computing management module and configure its own callable computing engine, while the data provider node as the provider can provide data sets and participate in computing through the privacy computing management module, and feed back the encrypted communication results to the demand side. The demand side can view the computing results on the central computing device, and the provider can view the computing process information in which it participates on the central computing device.

[0067] It should be noted that the data demand information may include the data demand node's requirements for data transaction types (such as computing models), data input and output requirements, etc.

[0068] Step S20: In response to the transaction order sent by the data providing node, task configuration information is generated based on the transaction order, and participating nodes are determined.

[0069] It should be noted that the transaction order may be the agreement information of the data transaction reached by the two parties. The above-mentioned participating nodes may be the data processing participating nodes participating in this data transaction among the blockchain member nodes, and the participating nodes include the data provider and the data participant.

[0070] It can be understood that the data demand node publishes data demand information through the central node, and the member nodes that meet the data demand conditions review the data demand information through the central node. After the review is passed, a transaction order is generated and fed back to the data demand node.

[0071] In some embodiments, the data demand of the data demand node may be a jointly constructed data model. The data demand node sends data demand information to the central node to define the calculation model, data input and output. The central node sends the data demand information to the blockchain. The data providing nodes in the member nodes that meet the conditions of the data demand information review the data demand information and the calculation model through the central node. After completion, a data transaction order is generated and fed back to the data demand node.

[0072] Furthermore, in order to accurately and reasonably configure the data processing task, the above step S20 may include:

[0073] In response to a transaction order sent by the data providing node, obtaining providing node information according to the transaction order;

[0074] Performing an interconnection test on the data providing node according to the providing node information;

[0075] In response to the interconnection test being successful, generating a data processing work order based on the transaction order;

[0076] Sending the data processing work order to the data providing node so that the data providing node authorizes the data processing work order;

[0077] In response to the data providing node agreeing to the authorization, task configuration information is generated based on the data processing work order, and participating nodes are determined.

[0078] It should be noted that the task configuration information can be the work order information generated by the data demand node based on the transaction order. The task configuration information can be used to configure the address and port of the participating nodes, the available data set catalog, the available privacy computing engine, and the computing mode (including hidden query, privacy intersection, multi-party secure computing, and federated modeling), and is distributed to each participating node involved in the calculation through the blockchain smart contract, including data providers and data participants.

[0079] In some embodiments, the data demander may generate work order information based on the transaction order, obtain data provider information based on the work order information, perform an interconnection test on the data provider node based on the data provider information, and after successfully connecting to the data provider node, add the other party to the partner list of this work order. The data provider authorizes the data to be used based on the work order information fed back by the data demander in order to provide data elements.

[0080] In some embodiments, the data processing flow is explained using joint modeling as an example: after the data demand node finds the data provider node according to the data processing work order and successfully connects, it applies for authorization for the data to be used. After successful authorization, the computing task is synchronized to all participating nodes.

[0081] The data provider node starts the modeling job locally based on its own privacy computing engine, and synchronizes the job status to the data demand node. The data demand node obtains the collected job parameters for modeling, and feeds back the global parameters to the data provider node in an encrypted manner. This iterative cycle continues until the model is built.

[0082] The final calculation results or training models are saved in the data demand node, and the data processing results can be published to the central node and uploaded to the portal homepage maintained by the central node.

[0083] Step S30: allocating data processing tasks to each participating node according to the task configuration information, wherein the data processing tasks include allocation parameters corresponding to each participating node.

[0084] It should be understood that the data demander generates task configuration information based on the work order information, creates data processing tasks based on the task configuration information, and specifies the allocation parameters of each participating node (the allocation parameters can be the specific privacy calculation parameters corresponding to each participating node in the data processing process). In addition, the job process of federated modeling is constructed using a visual orchestration tool and allocated to the data provider.

[0085] It should be noted that the above-mentioned data processing tasks may include the data set used for this task, the calculation mode of this task, and the data processing type (such as the type of model construction required for federated modeling).

[0086] Step S40: managing data asset rights based on the processing results of the data processing tasks fed back by each participating node according to the allocation parameters.

[0087] It should be noted that participating nodes (including data processing participating nodes and data providing nodes) and data demand nodes carry out privacy computing based on the assigned data processing tasks, such as using a universal base for communication and model parameter interaction of multi-platform privacy computing engines, and saving model instances and calculation results in the storage module of the demand-side device.

[0088] It should be noted that data assets may include raw data, analytical data, data instances and data products obtained through data processing (such as model instances).

[0089] It is understandable that the data demand node can determine the processing proportion of each participating node in this task based on the processing results of each participating node on the data processing task, thereby allocating and managing the use rights of the data assets of each participating node.

[0090] Furthermore, in order to accurately allocate the right to use the data assets to each participating node, the above step S40 may include:

[0091] Monitoring the processing status of the data processing task fed back by each participating node;

[0092] Based on the processing status, obtain the operation parameters fed back by each participating node;

[0093] Aggregate the job parameters fed back by each participating node to obtain the global parameters;

[0094] Encrypting the global parameter and sending the encrypted global parameter to the data providing node;

[0095] When the processing state is a processing completion state, determining a task processing weight of each participating node based on a processing result of the data processing task fed back by each participating node;

[0096] Allocating data usage rights to each participating node based on the task processing weight;

[0097] Manage data asset permissions based on data usage rights allocation results.

[0098] It can be understood that the participating nodes perform data processing jobs based on the data processing tasks assigned by the data demand nodes, and synchronize the job status to the data demand nodes. The data demand nodes obtain the aggregated job parameters for aggregation, obtain the global parameters, and feed back the global parameters to the data providing nodes in an encrypted manner. This iterative cycle continues until the data processing task is completed.

[0099] It should be noted that the data demand node synchronizes the processing status of the data processing tasks of each participating node to the central node. The central node determines the processing weight of each participating node for this data processing task based on the processing results of each participating node, thereby confirming the data asset permissions based on the processing weight and clarifying the allocation of usage rights of each participating node to the data asset permissions.

[0100] In this embodiment, data demand information is sent to the central node so that the central node determines the data providing node among the member nodes based on the data demand information, and sends the data demand information to the data providing node; in response to the transaction order sent by the data providing node, task configuration information is generated based on the transaction order, and participating nodes are determined; data processing tasks are allocated to each participating node according to the task configuration information, and the data processing tasks include allocation parameters corresponding to each participating node; data asset permissions are managed based on the processing results of the data processing tasks fed back by each participating node according to the allocation parameters; since this embodiment allocates data processing tasks to each member node in the blockchain through the blockchain central node, the management of data processing tasks is achieved by the central node, and the member nodes undertake their own data processing tasks according to the supply and demand parties, thereby ensuring efficient data management while effectively alleviating data processing pressure, and managing data asset permissions according to the processing results of each participating node, thereby achieving full life cycle management of data assets.

[0101] refer to Figure 3 , Figure 3 Schematic diagram of the second embodiment of the privacy data processing method of the present invention.

[0102] Based on the above first embodiment, in this embodiment, the privacy data processing method is applied to a participating node that participates in data processing among the member nodes of the blockchain, and the privacy data processing method includes:

[0103] Step S1: In response to a data processing task sent by a data demand node, an allocation parameter is obtained based on the data processing task.

[0104] It should be noted that the data processing task in this embodiment may be a joint modeling processing task or other federated data learning tasks. The above allocation parameters may be training parameters and model weight parameters of the training model.

[0105] It can be understood that this embodiment takes the joint modeling processing task as an example for explanation. It is assumed that the model weight W is iteratively trained in the federated learning task. In the tth round of iteration, the model weight of the data demand node is W t , each participating node obtains the allocation parameters assigned by the current model from the data demand node and performs multiple rounds of training locally. Assume that the participating node d i After multiple rounds of local training, the local model weights are obtained .

[0106] Step S2: Perform data processing based on the allocation parameters to obtain task processing parameters.

[0107] It should be noted that data processing can perform data processing tasks locally based on the training parameters allocated by the data demand node to obtain task processing parameters. For example, model training can be performed locally to obtain model weight parameters after multiple rounds of model parameter training.

[0108] Step S3: Encrypt the task processing parameters to obtain operation parameters.

[0109] It is understandable that the parameter node can generate a private key and a public key locally, encrypt data based on the private key and the public key, and obtain the operation parameters.

[0110] For example, the participating nodes use the public key to encrypt the local model weight parameter Wi to obtain the local encrypted model weight parameter .

[0111] Furthermore, in order to improve data security, the above step S3 may include:

[0112] Generate a local public key and obtain the public key of each participating node in the member node of the blockchain;

[0113] Aggregate the local public key and the public keys of the participating parties of each participating node to obtain an aggregated public key;

[0114] The task processing parameters are encrypted according to the aggregate public key to obtain operation parameters.

[0115] It is understandable that each participating node can generate a private key s locally i and public key b i , and aggregate the public keys bi of all participating nodes to calculate the aggregate public key , all participating nodes use the aggregated public key The task processing parameters calculated locally (such as the model weight parameter W i ) to encrypt the job parameters (such as local encryption model weight parameters ).

[0116] In some embodiments, all parameter nodes send local encryption model weights to data demand nodes for aggregation. For example, CKKS homomorphic encryption method can be used to support ciphertext addition. The data demand node obtains the encryption model weights and The data demand node will encrypt the model weights and parameters C sum1 Sent to each participating node, the participating nodes are based on C sum1 Calculate the local decrypted fragments and send them all back to the data demand node. The data demand node is based on the encryption model weight and C sum0 And the sum of the decrypted shards of all participating nodes Calculate the decrypted model weights and , after averaging, we get .

[0117] Combination Figure 4 The encryption calculation process is described below, assuming that the model weight w is iteratively trained in the federated learning task:

[0118] Step a: In the tth round of iteration, the model weight of the data demand node is w t , all participants obtain the global parameters assigned by the current model from the data demand node and perform multiple rounds of training locally. Assume that participant d i After multiple rounds of local training, the local model weights are obtained.

[0119] Step b: Each participating node generates a private key s locally i and public key b i , and aggregate the public keys bi of all participants to calculate the aggregate public key , all participants use the aggregate public key For the local model weight w i Encrypt to get the local encrypted model weight .

[0120] Step c: All parameter nodes send the local encryption model weights to the data demand node for aggregation. For example, the CKKS homomorphic encryption method can be used to support ciphertext addition. The data demand node obtains the encryption model weights and .

[0121] Step d: The data demand node will encrypt the model weights and parameters C sum1 Sent to each participating node, the participating nodes are based on Csum1 Calculate the local decryption fragment D i , and all are transmitted back to the data demand node.

[0122] Step e: Data demand nodes are based on the encrypted model weight and C sum0 And the sum of the decrypted shards of all participating nodes Calculate the decrypted model weights and , after averaging, we get If the model has converged at this time, save the model, otherwise repeat steps a to e until convergence.

[0123] Step S4: Send the job parameters to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node.

[0124] It can be understood that all participating nodes encrypt the task processing parameters after local data processing, and send the encrypted job parameters to the data demand node for aggregation. For example, each participating node can adopt the CKKS homomorphic encryption method for encryption, and the data demand node decrypts the job parameters based on the participating nodes to obtain the plaintext parameters, and obtains the processing results of the data processing task based on the plaintext parameters.

[0125] Furthermore, in order to ensure the integrity of the data task processing results and reduce the risk of data leakage, the above step S4 may include:

[0126] Sending the operation parameters to the data demand node, so that the data demand node aggregates the operation parameters sent by each participating node into a global parameter;

[0127] Obtaining global parameters sent by the data demand node;

[0128] Obtaining a decrypted segment based on the global parameter;

[0129] Adding noise information to the decrypted fragment based on a preset noise space to obtain a target decrypted fragment;

[0130] The target decrypted slice is sent to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the target decrypted slice sent by each participating node.

[0131] It can be understood that the CKKS algorithm is a homomorphic encryption scheme based on the RLWE (Ring Learning with Errors) assumption. When n is a power of 2 and q is a modulus, use represents the residual ring modulo q over the cyclotomic polynomial ring R, where is a polynomial ring with integer coefficients (modulo q). The basic assumption of RLWE is based on the parameter Given a polynomial pair , where a and b are uniformly randomly distributed on the ring, s is a ring element selected from the key distribution of the ring, and e is selected from the error distribution of the R ring. The ring elements selected above are described in detail below with specific steps:

[0132] Step 0: Given a security factor λ, set the dimension n, modulus q, and key distribution space based on the RLWE assumption , error distribution space and . Then generate a vector that follows a uniform random distribution , and returns the public parameter group Each participant d i Based on local key distribution Generate your own private keys i And calculate the corresponding public key . Then the aggregate public key is calculated based on the public keys provided by all participants as follows: .

[0133] Step 1: Local training model parameters. In the tth round of iterative training, the data demand node will set the global model weight parameter w t Distribute to all participants i , and the local model of the participant is trained based on local data. After multiple rounds of local iterations, the local model weight is obtained. .

[0134] Step 2: Encode and encrypt the model weights. To facilitate CKKS encryption, encode the weight vector into a plaintext polynomial to obtain the weight Plain text input After that, each participant d i In key distribution , Error distribution Sampling to get the key vector and the error vector . Based on this, calculate each participant d i The ciphertext model weight is as follows: .

[0135] In the above formula, After the calculation is completed, each participant d i The locally calculated ciphertext model weight Sent to the data request node.

[0136] Step 3: Homomorphic encryption calculates the weight and. The data demand node obtains the ciphertext model weights of all participants. After that, we can sum the ciphertext based on the principle of homomorphic encryption to get C sum As shown in the following formula, let , and C sum1 Distributed to all participants for calculating decrypted shard information, refer to the following formula: .

[0137] Step 4: Calculate the decrypted shard information. The ciphertext of the data demand node is equivalent to multi-party key encryption, so all participants are required to provide the decrypted shard D i To be able to fully decrypt, refer to the following formula: Here the participants need to decrypt the fragment D i Add noise to , The selection space is ,and The standard deviation of the distribution is greater than .

[0138] Step 5: Model aggregation. After the data demand node receives the decrypted shard Di from all participants, it combines C sum0 The model weights and decryption can be used to obtain the input plaintext , refer to the following formula: .

[0139] The data demand node obtains the decrypted model weight and input plaintext and obtains the model weight and , and then the average calculation is performed locally on the data demand node to obtain the new round of global model weights , refer to the following formula: .

[0140] Step 6: Model iteration. The data demand node calculates whether the model has converged based on the current state. If not, repeat steps 2 to 5. If it has converged, stop iteration and save the model results. This completes the distributed model training based on homomorphic encryption.

[0141] This embodiment responds to the data processing task sent by the data demand node, obtains allocation parameters based on the data processing task, performs data processing based on the allocation parameters, obtains task processing parameters, encrypts the task processing parameters, obtains job parameters, and sends the job parameters to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node, thereby enabling the participating nodes participating in the data processing to perform task operations on the allocated data processing tasks locally, and send the encrypted job parameters to the data demand node for encryption and aggregation, so that the data demand node obtains the processing result of the data processing task, effectively alleviating the data processing pressure while realizing data privacy processing, effectively avoiding the problem of data privacy leakage in the data processing process, and improving data security.

[0142] In addition, an embodiment of the present invention further proposes a computer-readable storage medium, on which a privacy data processing program is stored. When the privacy data processing program is executed by a processor, the steps of the privacy data processing method described above are implemented.

[0143] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0144] The computer-readable storage medium may be included in a data processing device, or may exist independently without being incorporated into a data processing device.

[0145] In addition, an embodiment of the present invention further provides a computer program product, including a privacy data processing program, which implements the steps of the privacy data processing method described above when executed by a processor.

[0146] The specific implementation methods of the computer program product of the present invention are basically the same as the embodiments of the above-mentioned privacy data processing method, and will not be repeated here.

[0147] Reference Figure 5 , Figure 5 This is a structural block diagram of the first embodiment of the privacy data processing device of the present invention.

[0148] like Figure 5 As shown, the privacy data processing device proposed in the embodiment of the present invention is used to implement the steps of the privacy data processing method as described above. The privacy data processing device can be applied to a trusted computing platform. The privacy data processing device can be composed of a central computing device and multiple member computing devices. The central computing device and the member computing devices can achieve functions such as intercommunication across government departments, consistent transaction management, and safe and feasible analysis and calculation through collaborative scheduling and blockchain evidence storage. Among them, the central computing device can be the central node of the blockchain, and the member computing device can be the member node of the blockchain.

[0149] It should be noted that the central computing device is mainly responsible for the functions of data circulation and computing scheduling, and it is necessary to deploy an authoritative department to support data management. The data mart centrally displays and conveniently queries the data resources published on the current platform. The transaction management module is responsible for matching the business of suppliers and demanders and forming orders. The privacy computing networking module is responsible for networking and configuring the member computing devices involved in the order. Finally, the key process of the transaction is realized through blockchain storage, and the ownership issues before and after data processing are clarified through smart contracts.

[0150] It should be noted that the member computing device mainly assumes two roles: provider and demander of privacy computing (wherein the provider can be a data provider node in the blockchain member node, and the demander can be a data demand node in the blockchain member node). The business department allocates roles according to the order requirements of data transactions, supports viewing the computing tasks initiated by its own server and participating in the computing tasks, and is mainly deployed in the data centers of various government commissions and bureaus. As a demander, you can initiate computing tasks through the privacy computing management module and configure your own callable computing engine, while as a provider, you can provide data sets and participate in computing through the privacy computing management module, and feedback the encrypted communication results to the demander. The demander can view the calculation results in the central computing device, and the provider can view the information of the participating computing process in the central computing device.

[0151] It is understandable that, referring to Figure 6 , Figure 6 This is a structural block diagram of the central computing device. The central computing device may be composed of multiple modules. The modules that constitute the central computing device may include a data mart module, a transaction management module, a privacy computing networking module and a blockchain evidence storage module.

[0152] The above-mentioned data mart module is used to provide a public resource directory of data sets to various government departments, can catalog and manage global data resources, and provide a portal homepage to support users to search for data sets or make relevant recommendations.

[0153] The above-mentioned transaction management module is used to support the management of system members and roles, support the issuance of data transaction orders to demanders and providers, and generate work orders based on the corresponding member computing device configuration and privacy computing configuration.

[0154] The above-mentioned privacy computing networking module is used to support the interconnection management of participants under the same work order, detect network connectivity, and support the configuration of member information, including address, port, computing task type, computing engine type and other information.

[0155] The above-mentioned blockchain evidence storage module is used to, on the one hand, provide reliable evidence for steps such as order formation and work order generation in the data transaction process, and ensure the security and non-tamperability of transactions through security audits; on the other hand, it checks the data authorization process through smart contracts to ensure that the ownership of the original data belongs to the data provider, and the ownership of the data calculation results belongs to the data demander.

[0156] It should be noted that the main function of the central computing device is to provide a platform for rapid matching between data demanders and data providers. First, the data demander determines the required data and creates an order on the portal homepage. The order clarifies the basic information and transaction information of the data supply and demand parties; then creates a corresponding work order based on the order information, including data sets, computing servers of the participating parties and other information; configures the node network through privacy computing networking management, adds the network information, port information and computing tasks of the participating parties, and then tests whether they are interoperable. If successful, it means that the work order configuration is normal and the work order can be issued. Finally, relying on the smart contract of the blockchain, the ownership of the original data and computing results of the transaction is clarified, and the transaction process is recorded.

[0157] It is understandable that, referring to Figure 7 , Figure 7 This is a structural block diagram of the member computing device. The main functions of the member computing device are composed of a privacy computing management module and a privacy computing engine module. Among them, the privacy computing management module includes a computing task management module, a computing data management module, a model visualization orchestration module, an engine configuration management module, and a blockchain evidence storage module.

[0158] The computing task management module is used to synchronize the work order information issued by the central computing device and construct the computing tasks of the member computing devices at this level, including data authorization, computing subtasks, computing models, and service publishing steps.

[0159] The computing data management module is used to manage the data sets involved in the computing tasks, including adding data sets, publishing data, viewing data details, and offline and deleting data sets.

[0160] The above-mentioned model visualization orchestration module is used to build and orchestrate the computing process when the computing task includes a federated learning task. Therefore, a visualization operator orchestration module is provided to edit operators online and support dragging modules to flexibly modify the computing business process.

[0161] The above-mentioned engine configuration management module is used to provide a management module for the privacy computing engine, supports registration of new computing engines, realizes engine access by configuring engine core ports and addresses and other information, and supports switching the computing engine type of the current server.

[0162] The above-mentioned blockchain evidence storage module is used by data providers to authorize demanders to provide data through smart contracts. At the same time, process information such as computing tasks and subtask management, computing engine service execution, task results and statistics are recorded in the blockchain to achieve safe and controllable data computing process.

[0163] It should be noted that the above-mentioned privacy computing engine module includes the existing privacy computing engine platform of the server at this level and a general privacy computing base module. The privacy computing engine platform can be subdivided into scheduling control, algorithm library and operator library according to the functions required by this system.

[0164] The above-mentioned universal privacy computing base module is used to provide a unified access base, which can connect to the scheduling control layer, algorithm layer, operator layer and other data information of multiple types of privacy computing platforms, and act as a standard medium for computing engines in the computing tasks of member computing devices.

[0165] The above-mentioned scheduling control is used for the scheduling control module of a single privacy computing platform, which supports users to select plaintext computing or ciphertext computing mode and provides control over computing tasks and algorithm scheduling.

[0166] The above algorithm library is used for the algorithm library of a single privacy computing platform. The mainstream privacy computing algorithms include privacy intersection (PSI), anonymous query, multi-party secure computing, and machine learning (logistic regression LR, neural network NN, and secure gradient boosting SecureBoost).

[0167] The above-mentioned operator library is used for the operator library of a single privacy computing platform, including secret sharing, oblivious transmission, obfuscated circuits, differential privacy, homomorphic encryption, etc.

[0168] It should be noted that the main function of the member computing device is to perform computing tasks. It is deployed in the data center nodes of various government departments and constitutes a government data element circulation network together with the central computing device through a private network. When the central computing device sends the work order to the member computing devices of the demander and the provider at the same time, and the demander is responsible for creating the computing task, the computing subtask can be constructed according to the supported computing types (including anonymous query, privacy intersection, multi-party secure computing, and joint modeling).

[0169] In some embodiments, joint modeling is used as an example to describe the joint modeling business process between the central computing device and the member computing devices as follows:

[0170] After the demander finds the provider based on the work order information and successfully connects with the provider, he / she applies for authorization for the data he / she needs to use. After successful authorization, the computing task is synchronized to all participating nodes. The provider starts the modeling job locally based on its own privacy computing engine and synchronizes the job status to the demander. The demander obtains the collected job parameters for modeling and feeds back the global parameters to the provider server in an encrypted manner. This iterative cycle continues until the model is built. The final calculation results or training model are saved on the demander's server, and the results can be published to the central computing device and put on the portal homepage.

[0171] It can be understood that the above-mentioned universal privacy computing base module can be a highly available privacy computing interconnection and interoperability standardized base. The universal privacy computing base module establishes unified standards, uses standardized protocols and standardized interfaces for connection, and realizes heterogeneous interactive collaboration at the three levels of data, project and algorithm protocols. It mainly includes two parts: a unified technology base and plug-in application services.

[0172] The above unified technical base is used for task scheduling and core management. The core functions are responsible for the implementation of the core algorithm capabilities of MPC (multi-party secure computing), PIR (hidden query), PSI (joint intersection) and FL (federated learning).

[0173] The above-mentioned plug-in application services are used to achieve interconnection and interoperability at the algorithm protocol layer and the algorithm data layer by applying the cryptographic algorithms and data standard protocols provided by a unified technology base, thereby breaking down computing silos and ensuring unimpeded flow of public data elements.

[0174] It should be noted that the general operation process of the privacy computing platform can be abstracted into four standard modules: scheduling control layer, computing execution layer, storage module and communication module. The general functions of the privacy computing platform mainly include scheduling control, algorithm library and operator library. By matching scheduling control with the process scheduling module, matching computing execution with the algorithm library and operator library, and providing standardized storage and data communication kits, these general components are encapsulated into the privacy computing standardized base. The privacy computing tasks between member computing devices communicate through the standard specifications and protocols of the base to achieve process coordination, computing coordination and data transmission. This technology can decouple the algorithms and operators of the privacy computing platform, thereby improving its interoperability and compatibility.

[0175] For example, refer to Figure 8 ,by Figure 8 Taking the structural block diagram of the privacy computing base module in China as an example, base A and base B achieve process coordination through the scheduling control layer, computing coordination through the computing execution layer, and data transmission through the communication module.

[0176] In this embodiment, data demand information is sent to the central node so that the central node determines the data providing node among the member nodes based on the data demand information, and sends the data demand information to the data providing node; in response to the transaction order sent by the data providing node, task configuration information is generated based on the transaction order, and participating nodes are determined; data processing tasks are allocated to each participating node according to the task configuration information, and the data processing tasks include allocation parameters corresponding to each participating node; data asset permissions are managed based on the processing results of the data processing tasks fed back by each participating node according to the allocation parameters; since this embodiment allocates data processing tasks to each member node in the blockchain through the blockchain central node, the management of data processing tasks is achieved by the central node, and the member nodes undertake their own data processing tasks according to the supply and demand parties, thereby ensuring efficient data management while effectively alleviating data processing pressure, and managing data asset permissions according to the processing results of each participating node, thereby achieving full life cycle management of data assets.

[0177] The privacy data processing device provided by the present application adopts the privacy data processing method in the above embodiment, which can solve the technical problem of data processing. Compared with the prior art, the beneficial effects of the privacy data processing device provided by the present application are the same as the beneficial effects of the privacy data processing method provided by the above embodiment, and the other technical features in the privacy data processing device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0178] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0179] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0180] In addition, for technical details not fully described in this embodiment, please refer to the privacy data processing method provided in any embodiment of the present invention, which will not be repeated here.

[0181] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0182] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0183] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0184] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A privacy data processing method, characterized in that: The privacy data processing method is applied to a data demand node with data demand among the member nodes of a blockchain, wherein the blockchain includes a central node and multiple member nodes, and the privacy data processing method includes: Sending data demand information to the central node, so that the central node determines a data providing node among the member nodes based on the data demand information, and sending the data demand information to the data providing node; In response to the transaction order sent by the data providing node, generating task configuration information based on the transaction order, and determining participating nodes; Allocate a data processing task to each participating node according to the task configuration information, wherein the data processing task includes allocation parameters corresponding to each participating node; wherein the allocation parameters are specific privacy calculation parameters corresponding to each participating node during the data processing process; The data asset authority is managed based on the processing result of the data processing task fed back by each participating node according to the allocation parameter.

2. The privacy data processing method according to claim 1, characterized in that: The step of responding to the transaction order sent by the data providing node, generating task configuration information based on the transaction order, and determining participating nodes includes: In response to a transaction order sent by the data providing node, obtaining providing node information according to the transaction order; Performing an interconnection test on the data providing node according to the providing node information; In response to the interconnection test being successful, generating a data processing work order based on the transaction order; Sending the data processing work order to the data providing node so that the data providing node authorizes the data processing work order; In response to the data providing node agreeing to the authorization, task configuration information is generated based on the data processing work order, and participating nodes are determined.

3. The privacy data processing method according to claim 1, characterized in that: The managing of data asset rights based on the processing result of the data processing task fed back by each participating node according to the allocation parameter includes: Monitoring the processing status of the data processing task fed back by each participating node; Based on the processing status, obtain the operation parameters fed back by each participating node; Aggregate the job parameters fed back by each participating node to obtain the global parameters; Encrypting the global parameter and sending the encrypted global parameter to the data providing node; When the processing state is a processing completion state, determining a task processing weight of each participating node based on a processing result of the data processing task fed back by each participating node; Allocating data usage rights to each participating node based on the task processing weight; Manage data asset permissions based on data usage rights allocation results.

4. A privacy data processing method, characterized in that: The privacy data processing method is applied to participating nodes that participate in data processing among member nodes of the blockchain, and the privacy data processing method includes: In response to a data processing task sent by a data demand node, an allocation parameter is obtained based on the data processing task; wherein the allocation parameter is a specific parameter of privacy calculation corresponding to each participating node in the data processing process; Perform data processing based on the allocation parameters to obtain task processing parameters; Encrypting the task processing parameters to obtain operation parameters; The job parameters are sent to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node.

5. The privacy data processing method according to claim 4, characterized in that: The step of encrypting the task processing parameters to obtain the operation parameters includes: Generate a local public key and obtain the public key of each participating node in the member node of the blockchain; Aggregate the local public key and the public keys of the participating parties of each participating node to obtain an aggregated public key; The task processing parameters are encrypted according to the aggregate public key to obtain operation parameters.

6. The privacy data processing method according to claim 5, characterized in that: The sending of the job parameters to the data demand node so that the data demand node obtains the processing result of the data processing task based on the job parameters sent by each participating node includes: Sending the operation parameters to the data demand node, so that the data demand node aggregates the operation parameters sent by each participating node into a global parameter; Obtaining global parameters sent by the data demand node; Obtaining a decrypted slice based on the global parameter; Adding noise information to the decrypted fragment based on a preset noise space to obtain a target decrypted fragment; The target decrypted slice is sent to the data demand node, so that the data demand node obtains the processing result of the data processing task based on the target decrypted slice sent by each participating node.

7. A privacy data processing device, characterized in that: The privacy data processing device is used to implement the steps of the privacy data processing method as described in any one of claims 1 to 3 or 4 to 6.

8. A data processing device, characterized in that: The data processing device includes: a memory, a processor, and a privacy data processing program stored in the memory and executable on the processor, wherein the privacy data processing program is configured to implement the privacy data processing method as described in any one of claims 1 to 3 or 4 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a privacy data processing program, and when the privacy data processing program is executed by the processor, the privacy data processing method according to any one of claims 1 to 3 or 4 to 6 is implemented.

10. A computer program product, characterized in that The computer program product comprises a privacy data processing program, which, when executed by a processor, implements the steps of the privacy data processing method according to any one of claims 1 to 3 or 4 to 6.

Images