Blockchain-based multi-party privacy computation method, device and system

Abstract

The application discloses a kind of multi-party privacy calculation method, device and system based on blockchain, and the method steps include: in task execution phase, task executor receives ciphertext proof sent by trusted third party, ciphertext proof is generated according to the data fingerprint generated by task executor and task publisher generated task ID, and chain is uploaded with task ID and ciphertext proof;In task result verification phase, task executor provides random sampling data to task publisher, to challenge randomly the data used in execution process, verify whether the execution data of task executor contains sampling data, in random challenge, the data fingerprint of the data participating in task by sampling in task publisher is sent to trusted third party with task ID, after obtaining ciphertext proof from blockchain, verify using challenge key to decrypt ciphertext proof.The application has the advantages of simple implementation method, high safety and reliability, verifiable correctness of task execution result, verifiable task work proof and the like.

Description

Technical Field

[0001] This invention relates to the field of privacy computing technology, and in particular to a blockchain-based multi-party privacy computing method, apparatus, and system. Background Technology

[0002] The following problems exist when using data: 1) Concerns about privacy data exposure; 2) Owners using fake data to perform tasks; 3) Data fraud is difficult to detect and prove. For example, the intermediate information exchanged in federated learning is highly condensed statistical information, lacking redundancy and interpretability, making it difficult to confirm whether these transmitted intermediate results are fabricated. Even if fraud is discovered, it is difficult to distinguish between maliciously fabricated data and noise caused by random sampling. Privacy-preserving computation (including multi-party secure computation, trusted computation, and federated learning) can be applied to scenarios where data is available but not visible, to solve the problems of data owners falsifying data and fraudulent computation tasks.

[0003] For privacy-preserving computation, existing technologies typically employ federated learning. The task publisher primarily verifies the task execution process using methods such as zero-knowledge proofs to generate proofs of the execution results and intermediate data, demonstrating the correctness of the task execution process without disclosing the executor's private data. The task publisher then verifies the proof, specifically verifying whether the result returned by the executor corresponds to the current task. However, this approach only verifies whether the result returned by the executor corresponds to the current task, without verifying the data used by the executor. Therefore, it cannot guarantee the authenticity of the data during task execution. For example, the executor might use malicious data or hide / filter sensitive data to execute the task, which would degrade the quality of the multi-party privacy-preserving computation results. Furthermore, the aforementioned methods cannot store the proof of task execution, thus making it impossible to trace and verify the data used in task execution. Summary of the Invention

[0004] The technical problem to be solved by this invention is: in view of the technical problems existing in the prior art, this invention provides a blockchain-based multi-party privacy computing method, device and system that is simple to implement, highly secure and reliable, verifiable in terms of the correctness of task execution results, verifiable in terms of task work proof, and traceable and verifiable in terms of the data used in task execution.

[0005] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows:

[0006] A blockchain-based multi-party privacy computation method includes the following steps:

[0007] During the task execution phase, the task executor receives a ciphertext certificate sent by a trusted third party. The ciphertext certificate is generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher, and the task ID and the ciphertext certificate are uploaded to the blockchain.

[0008] During the task result verification phase, the task executor provides randomly sampled data to the task issuer to randomly challenge the data used in the execution process and verify whether the task executor's execution data contains the sampled data. In the random challenge, the task issuer sends the sampled data fingerprint and task ID of the task participants to a trusted third party. After obtaining the encrypted proof from the blockchain, the third party uses the challenge key to decrypt the encrypted proof to achieve verification.

[0009] Furthermore, the steps for generating the ciphertext proof include:

[0010] A trusted third party assigns the master key and public key to the current task;

[0011] The task publisher generates a task ID for the current task;

[0012] The task executor uses its own private data to participate in the task, generates data fingerprints for the private data entries after the task is completed, and sends the data fingerprint set and task ID to a trusted third party;

[0013] A trusted third party generates encrypted proof based on the data fingerprint set and the task ID.

[0014] Furthermore, when the trusted third party allocates the master key and public key for the current task, it uses the initialization algorithm init(g,G0,α,β)→(pk,mk) to generate the public parameter public key pk and master key mk of the task, where G0 is a prime-order bilinear group, g is the generator of G0, and α and β are random exponents.

[0015] Furthermore, the task publisher generates a task ID for the current task by executing a task code generation algorithm, using the system timestamp datetime() and a random number random as parameters to generate a task code TaskID.

[0016] Furthermore, the step of generating the encrypted proof based on the data fingerprint set and the task ID includes:

[0017] Construct a proof-of-work strategy S based on the data fingerprint set finger_set and the task ID submitted by the task executor;

[0018] Based on the ciphertext generation algorithm, inputting common parameters pk and proof-of-work strategy S, generate ciphertext proof C:

[0019] Ecrypt CP-ABE(pk,S)→C, where Ecrypt CP-ABE () represents an encryption function using the CP-ABE encryption algorithm.

[0020] Furthermore, the proof-of-work strategy S is constructed according to the following formula:

[0021] {TaskID∧(finger1∨finger2...∨finger n ),n=sizeof(finger_set)}→S

[0022] Among them, finger i This represents the data fingerprint submitted by the i-th task executor, where i = 1, 2, ..., n, n is the number of data fingerprints, and TaskID represents the task code.

[0023] Furthermore, the steps for obtaining the challenge key include:

[0024] The task initiator samples a set of data fingerprints from the task executor and sends the task ID and the set of data fingerprints S to a trusted third party.

[0025] A trusted third party generates the current challenge key based on the master key mk, the task code TaskID, and the data fingerprint set S;

[0026] The task initiator receives the challenge key sent back by a trusted third party.

[0027] A blockchain-based multi-party privacy computing device includes:

[0028] The task execution control module is used to control the task executor to receive encrypted proof sent by a trusted third party during the task execution phase. The encrypted proof is generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher, and the task ID and encrypted proof are uploaded to the blockchain.

[0029] The task result verification control module is used to control the task executor to provide random sampled data to the task issuer during the task result verification stage, so as to randomly challenge the data used in the execution process and verify whether the execution data of the task executor contains the sampled data. In the random challenge, the task issuer sends the data fingerprint and task ID of the sampled participants to a trusted third party. After obtaining the encrypted proof from the blockchain, the third party uses the challenge key to decrypt the encrypted proof to achieve verification.

[0030] A blockchain-based multi-party privacy computing device includes a processor and a memory, wherein the memory stores a computer program and the processor executes the computer program to perform the method described above.

[0031] A blockchain-based multi-party privacy computing system includes a task initiator, a task executor, and a trusted third party for managing management keys and ciphertext, and further includes:

[0032] Blockchain is used to store the task ID and encrypted proof uploaded by the task executor;

[0033] The aforementioned blockchain-based multi-party privacy computing device is used to control data interaction between the task initiator, the task executor, the trusted third party, and the blockchain, thereby enabling multi-party privacy computing.

[0034] Compared with the prior art, the advantages of the present invention are as follows:

[0035] 1. This invention constructs a trust foundation for multi-party privacy computation proof of work using blockchain technology. It utilizes a partially centralized and tamper-proof blockchain to achieve trusted storage and verification of multi-party privacy computation proof of work. By fully leveraging the characteristics of blockchain, privacy computation can be achieved. This not only verifies the correctness of the execution result but also allows for the traceability and verification of the data used in the task execution. This makes the correctness of the task execution result verifiable, the task proof of work verifiable, and the data used in the task execution traceable and verifiable. Thus, it realizes the function of privacy computation: "data is available but not visible, data does not move but the model moves," which can greatly improve the security, reliability, and traceability of data.

[0036] 2. The present invention further utilizes a verifiable encryption method based on attribute-based encryption of data fingerprint sets to perform task proof ciphertext generation, challenge key generation, and challenge proof processes, thereby enabling lightweight verification of the work proof of the task executor. Attached Figure Description

[0037] Figure 1 This is a schematic diagram illustrating the implementation process of the blockchain-based multi-party privacy computation method in this embodiment.

[0038] Figure 2 This is a schematic diagram illustrating the structural principle of multi-party privacy computation based on blockchain in this embodiment.

[0039] Figure 3 This is a schematic diagram illustrating the detailed implementation process of ciphertext proof generation in this embodiment.

[0040] Figure 4 This is a detailed schematic diagram illustrating the implementation process of the multi-party privacy computation proof-of-work in this embodiment. Detailed Implementation

[0041] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0042] Blockchain is a chain-like database jointly built and maintained by computing nodes in a distributed peer-to-peer network, consisting of multiple linked data blocks. It features immutability, a decentralized organizational structure, and openness and privacy. This invention applies blockchain to achieve multi-party privacy computation. It uses blockchain technology to build a trust foundation for multi-party (task executor, task issuer, and trusted third party) privacy computation proofs of work. Through the decentralized and immutable nature of blockchain, it achieves trusted storage and verification of multi-party privacy computation proofs of work. By fully utilizing the characteristics of blockchain, privacy computation can not only verify the correctness of the execution result but also trace and verify the data used in task execution. This makes the correctness of the task execution result verifiable, the proof of work verifiable, and the data used in task execution traceable and verifiable. Thus, it achieves the function of "data usable but invisible, data immobile, model dynamic" in privacy computation, greatly improving data security, reliability, and traceability.

[0043] like Figure 1 As shown, the steps of the blockchain-based multi-party privacy computation method in this embodiment include:

[0044] S01. During the task execution phase, the task executor receives a ciphertext certificate sent by a trusted third party. The ciphertext certificate is generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher, and the task ID and the ciphertext certificate are uploaded to the blockchain.

[0045] S02. In the task result verification stage, the task executor provides randomly sampled data to the task issuer to randomly challenge the data used in the execution process, and verify whether the execution data of the task executor contains the sampled data. In the random challenge, the task issuer sends the fingerprint of the sampled data participating in the task and the task ID to a trusted third party. After obtaining the encrypted proof from the blockchain, the challenge key is used to decrypt the encrypted proof to achieve verification.

[0046] This embodiment generates a encrypted proof of the task executor's role in the current task based on the fingerprint of the data used in the task execution process and the task ID generated by the task publisher during the task execution phase. Then, during the task result verification phase, the task publisher randomly challenges the data used in the execution process. By randomly sampling the fingerprint and task ID of the data involved in the task, the system verifies whether the returned result belongs to the current task and whether the sampled data was used in the task execution process. This not only verifies the correctness of the task execution result but also verifies the proof of the task's work and allows for the traceability and verification of the data used in the task execution.

[0047] The specific steps for generating the ciphertext proof received in step S01 of this embodiment include:

[0048] S101. A trusted third party assigns a master key and a public key to the current task;

[0049] S102. The task publisher generates a task ID for the current task;

[0050] S103. The task executor uses its own privacy data to participate in the task, generates data fingerprints for privacy data entries after the task is completed, and sends the data fingerprint set and task ID to a trusted third party;

[0051] S104. A trusted third party generates a encrypted proof based on the data fingerprint set and the task ID.

[0052] In a specific application embodiment, when the aforementioned trusted third party allocates the master key and public key for the current task, it specifically uses the initialization algorithm init(g,G0,α,β)→(pk,mk) to generate the public parameter public key pk and master key mk for the task, where G0 is a prime-order bilinear group, g is the generator of G0, α and β are stochastic exponents, and α,β∈Z. p .

[0053] In a specific application embodiment, the task publisher generates a task ID for the current task by executing a task code generation algorithm, using the system timestamp datetime() and a random number random as parameters to generate a task code TaskID.

[0054] The specific steps in step S104 of this embodiment for generating the encrypted proof based on the data fingerprint set and the task ID include:

[0055] S141. Construct a proof-of-work strategy S based on the data fingerprint set finger_set and the task ID submitted by the task executor;

[0056] S142. Based on the ciphertext generation algorithm, input the common parameters pk and the proof-of-work strategy S to generate the ciphertext proof C:

[0057] Ecrypt CP-ABE (pk,S)→C, where Ecrypt CP-ABE () represents an encryption function using the CP-ABE encryption algorithm.

[0058] In a specific application embodiment, the above proof-of-work strategy S is constructed according to the following formula:

[0059] {TaskID∧(finger1∨finger2...∨finger n ),n=sizeof(finger_set)}→S (1)

[0060] Among them, finger iThis represents the data fingerprint submitted by the i-th task executor, where i = 1, 2, ..., n, n is the number of data fingerprints, and TaskID represents the task code.

[0061] In this embodiment, the specific steps for obtaining the challenge key used in step S02 include:

[0062] S201. The task initiator samples the data fingerprint set of the task executor and sends the task ID and the data fingerprint set S to a trusted third party;

[0063] S202. The challenge key for the current iteration is generated by a trusted third party based on the master key mk, the task code TaskID, and the data fingerprint set S;

[0064] S203. The task initiator receives the challenge key sent back by the trusted third party.

[0065] Through the above steps, a verifiable encryption method for data fingerprint sets based on attribute-based encryption technology can be realized. Combined with attribute-based encryption methods, it can effectively solve the problems of verifiable correctness of task execution results, verifiable proof of task work, and traceability of data used in task execution.

[0066] like Figure 2 As shown, this embodiment constructs a blockchain-based multi-party privacy computation system consisting of a task initiator, a task executor, a trusted third party, and a blockchain. The trusted third party is a fully trusted organization that manages keys and ciphertext, responsible for generating a master key and public key for the task initiator, and generating ciphertext proofs and challenge keys for the data fingerprint set of the task executor. The task executor is the owner of the privacy data and is responsible for using its own data to perform multi-party computation tasks. The task initiator is responsible for initiating multi-party privacy computation tasks and verifying the correctness of the task execution results and challenging the proof of work of the executor.

[0067] In specific application embodiments, such as Figure 2 As shown in the figure (the numbers in the figure correspond to the step numbers), the detailed steps of the above-mentioned blockchain-based multi-party privacy computation system using the multi-party privacy computation method of the present invention to implement multi-party privacy computation are as follows:

[0068] S01. Task Execution Process

[0069] Step (1) Task initialization: A trusted third party assigns a master key and a public key to the current task;

[0070] A trusted third party executes the initialization algorithm init(g,G0,α,β)→(pk,mk) to generate the public key and master key of the task, where G0 is a prime-order bilinear group and g is a generator of G0.

[0071] Step (2) Task code generation: The task publisher generates a task ID for the current task and sends the task ID to the task executor.

[0072] The task publisher generates the task ID by executing the task ID generation algorithm GenTaskID(datetime(),random)→TaskID. The task ID generation algorithm uses the system timestamp datetime() and the random number random as parameters, which can ensure that the task ID is unique and cannot be identified by the task executor in advance.

[0073] Step (3) The task executor uses its own privacy data to participate in the task, generates data fingerprints for privacy data entries after the task is completed, and sends the data fingerprint set and task ID to a trusted third party.

[0074] Step (4) Ciphertext Proof Generation: A trusted third party generates a working ciphertext proof for the data fingerprint set and task ID based on a verifiable encryption algorithm, and sends the ciphertext proof back to the task executor, such as... Figure 3 As shown.

[0075] The detailed steps for proving encrypted text include:

[0076] a. A trusted third party constructs a proof-of-work strategy S based on the data fingerprint set finger_set and the task ID submitted by the task executor. The specific construction method is shown in equation (1), namely:

[0077] {TaskID∧(finger1∨finger2...∨finger n ),n=sizeof(finger_set)}→S.

[0078] b. A trusted third party, based on the ciphertext generation algorithm, inputs public parameters pk and the proof-of-work strategy S, and outputs a ciphertext proof, namely Ecrypt. CP-ABE (pk,S)→C, where Ecrypt CP-ABE () represents an encryption function using the CP-ABE encryption algorithm.

[0079] Step (5) The task executor puts the task ID and encrypted proof on the blockchain.

[0080] Step (6) If it is trusted computing or federated learning, the task executor will send the execution result back to the task publisher, and the task publisher will process the returned result for the next step.

[0081] S02. Work Certificate Stage

[0082] like Figure 4 As shown, the work verification process includes:

[0083] Step (7) The task publisher randomly samples data from the task executor, and the task executor returns the fingerprint of the sampled data to the task publisher;

[0084] Step (8) The task issuer sends the task ID and sampled data fingerprint to a trusted third party, which generates a key for this random challenge based on the task ID and data fingerprint.

[0085] When generating the challenge key, the task initiator samples the data fingerprint set of the task executor and sends the task ID and data fingerprint set S to a trusted third party. The trusted third party then generates the challenge key based on the master key mk, the task code TaskID, and the data fingerprint set S.

[0086] KeyGen CP-ABE (mk,TaskID,S)→sk

[0087] A trusted third party sends the challenge key back to the task initiator.

[0088] Step (9) The task issuer obtains the ciphertext proof from the blockchain based on the task random number and decrypts the ciphertext proof using the challenge key. If the decryption is successful, it proves that the execution data of the task executor contains sampled data; otherwise, it proves that the execution data of the task executor does not contain sampled data.

[0089] After receiving the challenge key sk, the task initiator retrieves the working ciphertext proof of the task from the blockchain based on the TaskID, and decrypts the ciphertext C using the challenge key, i.e., Decrypt. CP-ABE (pk,C,sk)→m to determine whether the sampled data is included in the data used to execute the task.

[0090] This embodiment of the blockchain-based multi-party privacy computing device includes:

[0091] The task execution control module is used to control the task executor to receive encrypted proofs sent by a trusted third party during the task execution phase. The encrypted proofs are generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher, and the task ID and encrypted proofs are put on the blockchain.

[0092] The task result verification control module is used to control the task executor to provide random sampled data to the task issuer during the task result verification stage. This allows for random challenges to the data used in the execution process, verifying whether the task executor's execution data contains the sampled data. In the random challenge, the task issuer sends the fingerprint and task ID of the sampled data to a trusted third party. After obtaining the encrypted proof from the blockchain, the third party uses the challenge key to decrypt the encrypted proof and achieve verification.

[0093] In another embodiment, the blockchain-based multi-party privacy computing device of the present invention may further include a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to perform the blockchain-based multi-party privacy computing method as described above.

[0094] This embodiment is based on a blockchain-based multi-party privacy computation system, such as... Figure 2 As shown, it includes the task initiator, the task executor, and a trusted third party used to manage the management key and ciphertext, and also includes:

[0095] Blockchain is used to store the task ID and encrypted proof uploaded by the task executor;

[0096] The aforementioned blockchain-based multi-party privacy computing device is used to control the data interaction between the task initiator, the task executor, the trusted third party, and the blockchain to achieve multi-party privacy computing. The specific flowchart steps (1) to (9) are shown.

[0097] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A method for multi-party privacy computation based on blockchain, characterized in that the steps of include: During the task execution phase, the task executor receives a ciphertext certificate sent by a trusted third party. This ciphertext certificate is generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher. The task ID and the ciphertext certificate are then uploaded to the blockchain. Generating the ciphertext certificate based on the data fingerprint and task ID includes: Construct a proof-of-work strategy S based on the data fingerprint set finger_set and the task ID submitted by the task executor; Input the public key and public parameters according to the ciphertext generation algorithm. pk Work verification strategy S Generate ciphertext proof C : ,in Ecrypt CP-ABE ( ) indicates the use of CP-ABE The encryption function of the encryption algorithm; The proof-of-work strategy S is constructed according to the following formula: in, finger i This represents the data fingerprint submitted by the i-th task executor, where i = 1, 2, ..., n, and n is the number of data fingerprints. Indicates the task code; During the task result verification phase, the task executor provides randomly sampled data to the task issuer to randomly challenge the data used in the execution process and verify whether the task executor's execution data contains the sampled data. In the random challenge, the task issuer sends the sampled data fingerprint and task ID of the task participants to a trusted third party. After obtaining the encrypted proof from the blockchain, the third party uses the challenge key to decrypt the encrypted proof to achieve verification.

2. The blockchain-based multi-party privacy computation method according to claim 1, characterized in that, The steps for generating the ciphertext proof include: A trusted third party assigns the master key and public key to the current task; The task publisher generates a task ID for the current task; The task executor uses its own private data to participate in the task, generates data fingerprints for the private data entries after the task is completed, and sends the data fingerprint set and task ID to a trusted third party; A trusted third party generates encrypted proof based on the data fingerprint set and the task ID. 3.The method of claim 2, wherein, When the trusted third party allocates the master key and public key for the current task, an initialization algorithm is used. Public key of the generated task's public parameters pk and master key mk ,in It is a bilinear group of prime order. yes generator, and It is a stochastic index. 4.The method of claim 2, wherein, The task publisher generates a task ID for the current task, including executing a task code generation algorithm, taking the system timestamp datetime () and a random number random as parameters to generate the task code . 5.The method of any one of claims 1-4, wherein, The steps for obtaining the challenge key include: The task initiator samples a set of data fingerprints from the task executor and sends the task ID and the set of data fingerprints S to a trusted third party. A trusted third party generates a master key mk , a task code , and a set of data fingerprints S from the master key; and The task initiator receives the challenge key sent back by a trusted third party. 6.A blockchain-based multi-party privacy computing device, characterized in that, include: The task execution control module is used to control the task executor to receive encrypted proofs sent by a trusted third party during the task execution phase. These encrypted proofs are generated based on the data fingerprint generated by the task executor and the task ID generated by the task publisher. The task ID and the encrypted proof are then uploaded to the blockchain. Generating the encrypted proof based on the data fingerprint and task ID includes: Construct a proof-of-work strategy S based on the data fingerprint set finger_set and the task ID submitted by the task executor; Input the public key and public parameters according to the ciphertext generation algorithm. pk Work verification strategy S Generate ciphertext proof C : ,in Ecrypt CP-ABE ( ) indicates the use of CP-ABE The encryption function of the encryption algorithm; The proof-of-work strategy S is constructed according to the following formula: wherein, finger i represents the data fingerprint submitted by the i-th task executor, i = 1, 2, …, n, n is the number of data fingerprints, represents the task code; The task result verification control module is used to control the task executor to provide random sampled data to the task issuer during the task result verification stage, so as to randomly challenge the data used in the execution process and verify whether the execution data of the task executor contains the sampled data. In the random challenge, the task issuer sends the data fingerprint and task ID of the sampled participants to a trusted third party. After obtaining the encrypted proof from the blockchain, the third party uses the challenge key to decrypt the encrypted proof to achieve verification.

7. A blockchain-based multi-party privacy computation apparatus comprising a processor and a memory for storing a computer program, characterized in that, The processor is used to execute the computer program to perform the method as described in any one of claims 1 to 5.

8. A blockchain-based multi-party privacy computing system, comprising a task initiator, a task executor, and a trusted third party for managing management keys and ciphertext, characterized in that, Also includes: Blockchain is used to store the task ID and encrypted proof uploaded by the task executor; The blockchain-based multi-party privacy computing device as described in claim 6 or 7 is used to control data interaction between the task initiator, the task executor, the trusted third party, and the blockchain, thereby achieving multi-party privacy computing.

Images