Multi-party data cooperative computing method, electronic device, readable medium, program product, and participant device

Abstract

The invention provides a multi-party data cooperative computing method, electronic equipment, a readable medium, a program product and participant equipment, and the method comprises the steps: constructing a (k1)-degree polynomial with a constant term as original data in a finite field for the original data held by each participant, calculating function values of the polynomial at n different independent variable values to generate n data shares, wherein n is the number of participants; exchanging the data shares to enable each participant to obtain a data share set which is from all participants and corresponds to the same independent variable value; performing homomorphic cooperative calculation in a finite field based on the data share set to generate a calculation result share; k calculation result shares are collected, the shares are calculated in a finite field through an interpolation-based reconstruction method, and a final result of collaborative analysis is obtained. According to the invention, on the premise of not exposing the original data, the security cooperative calculation of the multi-party data can be realized.

Description

Technical Field

[0001] This disclosure relates to the fields of data security and privacy protection technology, and in particular to a multi-party data collaborative computing method, an electronic device, a computer-readable medium, a computer program product, and a participant device for multi-party data collaborative computing. Background Technology

[0002] In the era of big data, data has become a crucial production factor driving innovation and development across all sectors of society. The demand for cross-institutional and cross-sectoral collaborative data analysis is increasingly prominent, with applications spanning financial risk control, medical research, and government decision-making. Specifically, in the financial sector, collaborative analysis of data from multiple institutions can build more comprehensive customer profiles, thereby enabling more accurate risk assessment and supporting financial product innovation. In the medical field, integrating medical records and research data from different medical institutions for joint analysis helps accelerate disease research and improve clinical diagnosis and treatment outcomes. In government governance, aggregating data from various departments for collaborative analysis can provide data support for public policy formulation and urban governance.

[0003] However, as the value of data continues to increase, the security and privacy protection issues during data sharing and collaborative analysis have become increasingly prominent. Data breaches not only lead to the infringement of personal privacy and economic losses, but also expose related enterprises or institutions to legal and reputational risks. When national or public security data is involved, such issues are even more likely to affect national interests and social stability. Therefore, how to ensure data security and privacy while achieving efficient multi-party collaborative data analysis has become a crucial technical problem that urgently needs to be solved in this field. Summary of the Invention

[0004] This disclosure provides a multi-party collaborative computing method, an electronic device, a computer-readable medium, a computer program product, and a participant device for multi-party collaborative computing. The purpose of this disclosure is to achieve the correct acquisition of collaborative computing results without disclosing the original data of each participant, while reducing communication overhead and enhancing the participants' control over their own data, thereby addressing data security, efficiency, and trust issues in the prior art.

[0005] In a first aspect, embodiments of this disclosure provide a multi-party data collaborative computing method, comprising: constructing a constant term (k) of the original data held by each participating party within a finite domain. 1) A polynomial of degree n is generated, and the function values ​​of the polynomial at n different independent variable values ​​are calculated to generate n data shares, where n is the number of participants; the data shares are exchanged so that each participant obtains a set of data shares corresponding to the same independent variable value from all participants; homomorphic collaborative computation is performed on the data share set in the finite domain to generate computation result shares; k computation result shares are collected, and the shares are calculated in the finite domain using an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; wherein the final result is consistent with the result of directly performing the homomorphic collaborative computation on each original data.

[0006] In some embodiments, the (k) 1) The coefficients of the polynomial, excluding the constant term, are generated by random selection within the finite field, so that fewer than k data shares cannot be used to deduce the corresponding original data.

[0007] In some embodiments, the modulus of the finite field is greater than the sum of the original data of all participants.

[0008] In some embodiments, exchanging the data shares includes: encrypting the data shares using an encryption algorithm; and transmitting the encrypted data shares to the corresponding participants via a secure network transmission protocol.

[0009] In some embodiments, before performing homomorphic collaborative computation within the finite domain based on the set of data shares, the method includes: decrypting the encrypted data shares after receiving them.

[0010] In some embodiments, the homomorphic collaborative computation includes performing an addition operation on the set of data shares without restoring the original data.

[0011] In some embodiments, the multi-party data collaborative computation method further includes: calculating the shares based on an interpolation-based reconstruction method within the finite domain using another set of k different computation result shares to obtain a verification result; determining whether the verification result is consistent with the final result, and if they are consistent, confirming that the final result is correct.

[0012] In a second aspect, embodiments of this disclosure provide an electronic device, including: one or more processors; and a memory storing one or more programs thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the multi-party data collaborative computing method described in the first aspect of this disclosure.

[0013] Thirdly, embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the multi-party data collaborative computing method described in the first aspect of embodiments of this disclosure.

[0014] Fourthly, embodiments of this disclosure provide a computer program product, including a computer program or instructions, which, when executed by a processor, implement the multi-party data collaborative computing method described in the first aspect of this disclosure.

[0015] Fifthly, embodiments of this disclosure provide a participant device for multi-party collaborative data computation, including a processor and a memory, wherein the processor is configured to execute a program stored in the memory, such that the participant device is able to: construct a constant term (k) of the original data within a finite domain for the original data it holds. 1) A polynomial of degree n is used to calculate the function values ​​of the polynomial at n different independent variable values ​​to generate n data shares, where n is the number of participants; the data shares are exchanged with other participating devices so that each participating device obtains a set of data shares corresponding to the same independent variable value from all participating devices; homomorphic collaborative computation is performed on the data share set in the finite domain to generate computation result shares; as a requester, the computation result shares are obtained from other participating devices, and when k computation result shares are obtained, the computation result shares are calculated in the finite domain using an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; and as a responder, in response to the requests of other participating devices, the computation result shares are provided to the requester.

[0016] This disclosure enables secure data sharing and collaborative computation among multiple participants without disclosing the original data, and ensures the correctness and verifiability of the computation results. Attached Figure Description

[0017] Figure 1 This is a flowchart of a multi-party data collaborative computing method according to an embodiment of this disclosure.

[0018] Figure 2 This is a block diagram of a multi-party data collaborative computing system according to an embodiment of the present disclosure.

[0019] Figure 3 This is a schematic diagram of the composition of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions of this disclosure will be described in detail below with reference to the accompanying drawings.

[0021] Exemplary embodiments will be described more fully below with reference to the accompanying drawings; however, these exemplary embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will enable those skilled in the art to fully understand the scope of this disclosure.

[0022] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0023] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0024] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. As used herein, the singular forms “a” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded.

[0025] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.

[0026] This disclosure relates to a system and method for multi-party secure collaborative data analysis based on a secret sharing algorithm. It can be applied to scenarios where joint data analysis is carried out while ensuring the data security and privacy of all parties, such as joint risk assessment between financial institutions, collaborative analysis of cross-institutional medical and scientific research data, and data-driven collaborative decision-making between government departments.

[0027] This disclosure aims to address the common technical problems in existing multi-party collaborative data analysis technologies, such as high data security risks, high communication costs, and limited willingness of participating parties to collaborate. To this end, this disclosure introduces a threshold secret sharing mechanism (e.g., the Shamir secret sharing algorithm) to secretly divide the original data held by each participating party, allowing the original data to participate in subsequent calculations in the form of data shares. This achieves data "usable but not visible" throughout the entire collaborative analysis process, avoiding the security risks associated with centralized storage or plaintext exchange of original data, and fundamentally reducing the risk of data leakage.

[0028] Meanwhile, this disclosure adopts a distributed collaborative computing architecture, where each participant only needs to perform a preset homomorphic collaborative computing on its own data share locally, without having to upload the original data or complete intermediate computing results to the central node or a third-party platform, thereby significantly reducing the data transmission scale between the participants and the central platform and reducing the overall communication overhead.

[0029] Furthermore, each participant only holds a data share corresponding to its own identity and cannot independently recover the original data or global data view of any other participant. This effectively enhances the participants' control over their own data and the foundation of trust, thereby increasing their enthusiasm for participating in multi-party data collaborative analysis and improving the overall scalability of the system.

[0030] Figure 1 This is a flowchart of a multi-party data collaborative computing method according to an embodiment of this disclosure.

[0031] Firstly, referring to Figure 1 This disclosure provides a multi-party data collaborative computing method, including: S11. For the original data held by each participant, construct a constant term within a finite field that is (k) of the original data. 1) A polynomial of degree n is used to generate n data shares, where n is the number of participants. S12. Exchange the data shares so that each participant obtains a set of data shares from all participants that correspond to the same independent variable value; S13. Based on the data share set, perform homomorphic collaborative computation within the finite domain to generate computation result shares; S14. Collect k computation result shares, and calculate the shares in the finite domain using an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; wherein the final result is consistent with the result of directly performing the homomorphic collaborative computation on each original data.

[0032] In some embodiments, the (k) 1) The coefficients of the polynomial, excluding the constant term, are generated by random selection within the finite field, so that fewer than k data shares cannot be used to deduce the corresponding original data.

[0033] In some embodiments, the modulus of the finite field is greater than the sum of the original data of all participants.

[0034] In this embodiment, the secret sharing algorithm can be implemented based on the Shamir secret sharing algorithm or other polynomial-type secret sharing algorithms with equivalent threshold security characteristics. In such algorithms, the secret is encoded as a constant term of a polynomial, and shares are generated by calculating the values ​​of this polynomial at different points.

[0035] In this embodiment of the disclosure, during the data segmentation step, each participant (or data holder) independently processes the original data they hold. The core of this step lies in using threshold secret sharing technology to securely transform the original data into multiple data shares that can be held in a decentralized manner.

[0036] Specifically, multi-party data collaborative computing is achieved through the following process: First, a suitable large prime number q is chosen to construct a finite field GF(q). The selection of the modulus q of this finite field must satisfy a key condition: it must be greater than the possible numerical range of the original data of all participants. In a preferred embodiment, when the computational operation of the collaborative analysis is addition (e.g., calculating a sum), the modulus q is configured to be greater than the sum of the original data of all participants.

[0037] Within a selected finite field GF(q), each participant constructs a random polynomial for its original secret value S (i.e., the original data to be protected). This polynomial is k. Once, its general form is: f(x) = a0 + a1x + a2x² + ... + a k 1x k 1 , In this polynomial, the constant term a0 is directly set to the original data S. The other coefficients besides the constant term are a1, a2, ..., a... k 1. All data shares are randomly selected within the finite field GF(q). This randomization process ensures that, according to the mathematical principle of secret sharing, no combination of fewer than k data shares can derive any information about the original data S, thus achieving security in the information theory sense and meeting the data privacy requirements of this system.

[0038] Suppose there are n participants (n≥k). Each participant calculates the function values ​​of the polynomial f(x) at n distinct, pre-agreed values ​​of the independent variable x1, x2, ..., xn. Typically, for simplicity and standardization, xi = i (i = 1, 2, ..., n), for example, corresponding to the participant's number or a pre-agreed index value. For each i, calculate: si=f(i) mod q This yields n data shares s1, s2, ..., sn. These shares constitute the secret partitioning of the original data S. Each participant generates a set of data shares corresponding to the n values, which are used for subsequent secure distribution and processing.

[0039] In some embodiments, exchanging the data shares includes: encrypting the data shares using an encryption algorithm; and transmitting the encrypted data shares to the corresponding participants via a secure network transmission protocol.

[0040] In this embodiment, during the transmission step, to ensure the confidentiality and integrity of data shares during network transmission between participants and to prevent eavesdropping or tampering of shares in the transmission link, share encryption can be performed. Before sending its data shares to the target participant, the sending participant first encrypts each share to be transmitted. To improve encryption / decryption efficiency and ensure strength, a symmetric encryption algorithm, such as the Advanced Encryption Standard (AES) algorithm, is preferably used. The encrypted share ciphertext is sent through a network transmission protocol with secure communication capabilities; in this embodiment, Hypertext Transfer Security Protocol (HTTPS) is preferably used. Through the HTTPS channel, the encrypted shares are reliably transmitted to the target participant. The receiving participant receives the encrypted data shares through a secure channel and decrypts the ciphertext in subsequent processing to obtain the corresponding data share plaintext, which is then added to its local data share set. Encryption enables the secure movement of data shares from the sender to the receiver, thereby completing data transmission between participants without exposing the content of the data shares.

[0041] In some embodiments, before performing homomorphic collaborative computation within the finite domain based on the set of data shares, the method includes: decrypting the encrypted data shares after receiving them.

[0042] In some embodiments, the homomorphic collaborative computation includes performing an addition operation on the set of data shares without restoring the original data.

[0043] In this embodiment of the disclosure, in the collaborative computing step among the participants, each participant, without restoring any original data, collaborates to complete the predetermined computing task based solely on the set of secret shares held locally, and generates the computing result share required for the next step.

[0044] After receiving the encrypted data share, each participant first decrypts it to recover the data share. After locally possessing a complete set of data shares from all participants corresponding to the same independent variable value, each participant independently performs a pre-defined homomorphic collaborative computation within a selected finite field GF(q) based on its own set of data shares. Directly performing this homomorphic collaborative computation on the secretly shared data shares results in a share that, after subsequent reconstruction steps, is mathematically equivalent to the result obtained by directly performing the same computation on the original plaintext data of all participants. This achieves "data usable but invisible." In a preferred embodiment, the homomorphic collaborative computation is specifically an addition operation. Each participant sums all shares in its local data share set (i.e., shares from each participant at the same independent variable value) within the finite field GF(q). That is, for a participant with identity identifier (independent variable value) j, the calculation is: Sum j = mod q Where s i,j This represents the share generated and distributed by the i-th participant to the j-th participant. The calculated Sum... j This refers to the share of the computational result generated by that participant. Throughout the computation process, each participant only operates on secret shares and cannot, and does not need to, recover any other participant's original data from these shares. The computation is completed entirely locally, without interacting with original data or exposing intermediate computation states during the computation process, thus maximizing data privacy. After generating the computational result share, the calculated result share is encrypted and sent to the subsequent result aggregation step.

[0045] In some embodiments, the multi-party data collaborative computation method further includes: calculating the shares based on an interpolation-based reconstruction method within the finite domain using another set of k different computation result shares to obtain a verification result; determining whether the verification result is consistent with the final result, and if they are consistent, confirming that the final result is correct.

[0046] In this embodiment of the disclosure, in the result summarization step, the share of the calculation results from a sufficient number of participants is collected, the final plaintext result of the collaborative analysis is recovered through mathematical reconstruction, and the result is optionally verified to enhance credibility.

[0047] Any participant requiring computation results (as the "requesting party") or a designated aggregation node initiates a request to other participants to collect their respective shares of the computation results. Collecting any k shares of computation results from different participants allows for subsequent reconstruction. After collecting k shares of computation results, computation is performed within the previously selected finite field GF(q) using an interpolation-based reconstruction method. In a preferred embodiment, Lagrange interpolation is employed.

[0048] Let the collected k shares be (x1, y1), (x2, y2), ..., (xk, yk), where xr is the identity identifier of the share provider (the value of the independent variable), and yr is the share of the calculated result provided by the provider. Reconstruct the polynomial F(x) of the calculated result using the Lagrange interpolation formula, and calculate the value of this polynomial at the zero point x=0: The calculated F(0) is the final result of the collaborative analysis.

[0049] Based on the homomorphic property of secret sharing, the final computation result F(0) recovered by the above reconstruction process is mathematically equivalent to the result obtained by hypothetically performing homomorphic collaborative computation directly on the original data of all participants. The final result can be output, stored, or used for subsequent business processing.

[0050] To further enhance the system's robustness and resist potential computational errors or malicious behavior, this method may include an optional verification step. After obtaining the final result, a verification result can be obtained by recalculating within the finite domain using the same interpolation-based reconstruction method based on another set of k different computation result shares (i.e., the share set not used in the above reconstruction), and then determining whether the verification result is consistent with the previously obtained final result. If they are consistent, the final result can be confirmed to be correct with a very high probability. This is because both independent sets of shares originate from the same batch of correct local computations, and the same mathematical reconstruction should yield a unique result. If they are inconsistent, it indicates that at least one set of shares or an error exists in the computation process, at which point the system can trigger an alarm, request more shares for recalculation, or conduct an audit. This verification step, as an optional function to enhance security and reliability, does not constitute a necessary step in the multi-party data collaborative computation method, but it provides important added value to this system in application scenarios requiring high reliability.

[0051] Figure 2 This is a block diagram of a multi-party data collaborative computing system according to an embodiment of the present disclosure.

[0052] Secondly, referring to Figure 2This disclosure provides a multi-party data collaborative computing system, including multiple participating node nodes, each of which includes: Data segmentation module 201 is configured to construct a constant term (k) of the original data within a finite domain for the original data it holds. 1) A polynomial of degree n is used to generate n data shares, where n is the number of participants. Transmission module 202 is configured to exchange the data shares among participating nodes, so that each participating node obtains a set of data shares corresponding to the same independent variable value from all participating nodes; The collaborative computing module 203 is configured to perform homomorphic collaborative computing within the finite domain based on the data share set, and generate computing result shares; The result aggregation module 204 is configured to collect k computation result shares, and calculate the shares in the finite domain using an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; wherein the final result is consistent with the result of directly performing the homomorphic collaborative computation on each original data.

[0053] In this embodiment, the multi-party data collaborative computing system adopts a distributed and modular design, with its core consisting of multiple functionally equivalent participant nodes. Each participant node is deployed locally on the local machine of its respective data holder and includes a data segmentation module, a transmission module, a collaborative computing module, and a result aggregation module. These modules interact through well-defined secure interfaces, ensuring that the entire collaborative analysis process operates securely and efficiently under end-to-end data encryption protection.

[0054] The data segmentation module is responsible for receiving the original data (i.e., the secret value) held by this node and securely segmenting it into multiple data shares according to the threshold secret sharing principle (such as the Shamir secret sharing algorithm). The module operates within a finite domain with a sufficiently large modulus. It constructs a (k-1)th degree random polynomial with a constant term for the original data and calculates the function values ​​of this polynomial at n different and pre-agreed values ​​of the independent variable (e.g., values ​​that are unique identifiers for each node). The resulting n function values ​​are the n data shares of the original data.

[0055] The transmission module is responsible for securely managing the distribution and reception of data shares across the network. The module receives a share list from the data segmentation module. Before transmission, to ensure security, it encrypts each share to be sent. In a preferred embodiment, symmetric encryption algorithms such as Advanced Encryption Standard (AES) are used for encryption, and the encrypted shares are accurately sent to the corresponding participating nodes via secure network transmission protocols such as Hypertext Transfer Security (HTTPS). Simultaneously, the module also receives encrypted shares from other participating nodes. After decryption, it organizes the data shares from all participating nodes (including itself) corresponding to the same independent variable value into a locally held set of data shares.

[0056] The collaborative computing module performs a pre-defined homomorphic collaborative computing task within the finite domain based on the local data share set provided by the transmission module. The core characteristic of this computation is "homomorphism," meaning that directly computing on the secret share yields a result share that, after subsequent reconstruction, is completely consistent with the final result obtained by directly computing all the original plaintext data. This ensures that "data is usable but not visible." In a typical embodiment, this computation is an addition operation; the module sums all shares in the local data share set within the finite domain to generate a single result share. It is important to emphasize that the module performs the computation only after the transmission module has decrypted all received shares, enabling operations based on the plaintext shares; however, the entire process never attempts to recover the original data of any other participant.

[0057] The results aggregation module is responsible for collecting intermediate computation results from multiple participating nodes and reconstructing the final plaintext result of the collaborative analysis. When any node needs to obtain the result, its local results aggregation module can be activated. It requests and collects k distinct shares of the computation result from other nodes. Within the finite domain, this module uses an interpolation-based reconstruction method (e.g., Lagrange interpolation) to calculate these shares and recover the final result of the collaborative analysis. Based on the principle of homomorphism, this final result is guaranteed to be consistent with the result of directly performing the aforementioned homomorphic collaborative computation on the original data of each participating node.

[0058] To enhance reliability, the result aggregation module can also be configured to perform result verification and auditing. For example, it can be reconstructed based on another set of k different calculation result shares, and the correctness of the final result can be verified by comparing whether the two reconstruction results are consistent. The above result verification and auditing functions are only an optional implementation method to enhance reliability and do not constitute a necessary limitation on the multi-party data collaborative computing system.

[0059] The system's modules interact through secure interfaces, and data is encrypted during transmission between different modules, ensuring data security and privacy throughout the system's operation.

[0060] Figure 3 This is a schematic diagram of the composition of an electronic device according to an embodiment of the present disclosure.

[0061] Thirdly, referring to Figure 3 This disclosure provides an electronic device, which includes: One or more processors 301; The memory 302 stores one or more programs, which, when executed by one or more processors, enable the one or more processors to implement any of the above-mentioned multi-party data collaborative computing methods. One or more I / O interfaces 303 are connected between the processor and the memory and configured to enable information exchange between the processor and the memory.

[0062] Among them, processor 301 is a device with data processing capabilities, including but not limited to central processing unit (CPU); memory 302 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH); I / O interface (read-write interface) 303 is connected between processor 301 and memory 302, and can realize information interaction between processor 301 and memory 302, including but not limited to data bus (Bus).

[0063] In some embodiments, the processor 301, memory 302, and I / O interface 303 are interconnected via bus 304, and thus connected to other components of the computing device.

[0064] Fourthly, embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described multi-party data collaborative computing methods.

[0065] Fifthly, embodiments of this disclosure provide a computer program product, including a computer program or instructions, which, when executed by a processor, implement any of the aforementioned multi-party data collaborative computing methods.

[0066] Sixthly, embodiments of this disclosure provide a participant device for multi-party collaborative data computing, including a processor and a memory, wherein the processor is configured to execute a program stored in the memory, such that the participant device is capable of: For the original data it holds, construct a constant term within a finite field that is (k) of the original data. 1) A polynomial of degree n is generated, and the function values ​​of the polynomial at n different independent variable values ​​are calculated to generate n data shares, where n is the number of participants; The data shares are exchanged with other participating devices so that each participating device obtains a set of data shares corresponding to the same independent variable value from all participating devices; Based on the data share set, homomorphic collaborative computation is performed within the finite domain to generate computation result shares; As the requesting party, it obtains the computation result shares from the devices of other participating parties, and when it obtains k computation result shares, it calculates the computation result shares in the finite domain through an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; And as a respondent, in response to a request from another participating device, it provides the requesting party with its own share of the computational results.

[0067] The key technical features of this disclosure are mainly reflected in the following aspects: (1) Threshold-based secret sharing of raw data segmentation: In a finite domain, the raw data of each participant is encoded into the constant term of a polynomial, and multiple data shares are generated through a random polynomial, so that the raw data exists among multiple participants in the form of distributed shares. The "raw data" is transformed from the system structure level into irreversible, distributed mathematical shares, ensuring data privacy from the source.

[0068] (2) Correctness guarantee design based on finite field constraints: By selecting a finite field whose modulus is greater than the expected range of the calculation result, the arithmetic correctness of collaborative computation is embedded as a hard constraint into the system foundation to ensure that the calculation performed on the share is mathematically equivalent to the calculation on the original plaintext data, thereby avoiding computational overflow or result ambiguity.

[0069] (3) Decentralized peer-to-peer share exchange and local holding architecture: Each participant securely exchanges shares through peer-to-peer means, so that each node only holds a set of shares from all nodes corresponding to its own identifier. There is no central node at the system level that centrally stores or processes all data (or shares), structurally eliminating the "data central point" and reducing single point leakage and trust risks from the architecture.

[0070] (4) Transmission mechanism combining mathematical security and communication security: During the data share exchange process, the data shares are encrypted and transmitted through a secure communication protocol, thereby maintaining the security of secret sharing of mathematics while further preventing the data shares from being eavesdropped on or tampered with during network transmission. This combines cryptographic security (secret sharing) with engineering security (secure transmission).

[0071] (5) Privacy-preserving computation engine based on share homomorphism: Each participant directly performs homomorphic computation operations (such as addition) within a limited domain based on its locally held share set to generate the calculated share. The entire computation process does not require the recovery of any original data, realizing a computation paradigm of "data available but not visible".

[0072] (6) Threshold-based reconstruction result recovery mechanism: By collecting any number of computation result shares that meet the threshold, and using an interpolation-based reconstruction method within the same finite domain, the final result of collaborative computation is recovered, thereby achieving the correct output of distributed computation results. Result recovery does not depend on specific nodes or all participants, and has fault tolerance and failure resistance capabilities.

[0073] (7) Optional distributed cross-validation mechanism: It can independently reconstruct based on another set of different result shares that meet the threshold, and compare the two reconstruction results to mathematically verify the correctness of the final result, providing the system with the ability to resist malicious behavior and enhance credibility.

[0074] By implementing the above technical solution, this disclosure has achieved significant beneficial effects, specifically reflected in the following three aspects: (1) In terms of data security and privacy protection, this disclosure achieves fundamental privacy protection and enhanced transfer security. First, by transforming the original data into a threshold secret sharing form, the data exists in the form of invisible mathematical shares throughout its entire lifecycle, ensuring from a mathematical mechanism perspective that even if some nodes or communication links of the system are compromised, the original data cannot be recovered. Second, on this basis, an encrypted transmission mechanism is superimposed, effectively protecting the shares from risks during network transmission and improving the overall security of the system in actual deployment environments.

[0075] (2) At the level of computational correctness and system architecture, this disclosure ensures the mathematical reliability of the computation results and constructs a decentralized, highly reliable architecture. On the one hand, through the design of constraints such as finite fields, it ensures that the collaborative computation results are mathematically consistent with the results of directly computing the original data, thus avoiding distortion. On the other hand, by adopting a decentralized share holding and computation model, it completely eliminates the dependence on centralized nodes or trusted third parties, reducing the risk of single point of failure and centralized attacks from the perspective of system architecture. On this basis, the system only needs to collect shares that meet the threshold to output the results, which naturally has fault tolerance and can flexibly cope with the situation of partial node failure.

[0076] (3) In terms of collaborative analysis utility and trusted verification, this disclosure achieves true "data usability without visibility" and provides enhanced assurance of result credibility. Its core lies in the fact that the system can directly perform homomorphic computation on data in a secret shared form, thereby completing collaborative analysis tasks without exposing the original data, achieving an effective balance between privacy protection and data availability. Furthermore, by introducing an optional independent reconstruction verification mechanism, the system can effectively detect errors or malicious behavior, significantly enhancing the credibility and auditability of the final results, making it suitable for scenarios with extremely high requirements for result accuracy.

[0077] To enable those skilled in the art to more clearly understand the technical solutions provided by the embodiments of this disclosure, the technical solutions provided by the embodiments of this disclosure will be described in detail below through specific embodiments: Example 1: Multi-party collaborative calculation of a depositor's total deposits involving 5 banks based on the Shamir algorithm. This embodiment uses the scenario of five banks (A, B, C, D, and E) collaboratively calculating a depositor's total deposits without disclosing their individual deposit amounts to demonstrate the implementation process of the disclosed method. The system parameters are set as follows: number of participants n=5, threshold value k=3. The estimated total deposits are no more than 4 million yuan, and a finite field modulus q=503 (satisfying q>400) is selected.

[0078] Step 1: Data Segmentation and Share Generation (Each bank generates its own secret share of deposits) Each bank uses its own deposit amount as a secret value S, and independently constructs it twice (i.e., k) within the finite field GF(503). The first-order random polynomial f(x) = S + a1x + a2x 2 The coefficients a1 and a2 are randomly selected from GF(503).

[0079] Each bank calculates the function value f(i) mod 503 of the polynomial at x = 1, 2, 3, 4, 5 (corresponding to the identity identifiers of banks A to E respectively), thereby generating 5 data shares for its own deposits. For example: Bank A (S) A =500,000 yuan, randomly select coefficients a1=12, a2=25 (random integers within GF(503)) to generate shares: A1=87, A2=174, A3=311, A4=498, A5=232.

[0080] Bank B (S) B =800,000 yuan, randomly select b1=18, b2=30) to generate shares: B1=128, B2=236, B3=404, B4=129, B5=417.

[0081] Bank C(S) C =1.2 million yuan, randomly select c1=22, c2=15) to generate shares: C1=157, C2=224, C3=321, C4=448, C5=102.

[0082] Bank D (S) D =600,000 yuan, randomly select d1=9, d2=20) to generate shares: D1=89, D2=158, D3=267, D4=416, D5=102.

[0083] Bank E (S) E =900,000 yuan, randomly select e1=15, e2=10) to generate shares: E1=115, E2=160, E3=225, E4=310, E5=415.

[0084] The amounts and units mentioned above are only used to illustrate the calculation process and do not constitute a limitation on the actual numerical range.

[0085] Step 2: Secure Distribution and Aggregation of Shares Each bank will encrypt and send its five generated shares to five banks A, B, C, D, and E respectively (for example, bank A sends A1 to itself, A2 to bank B, A3 to bank C, and so on).

[0086] Each bank digitally signs (e.g., RSA signature) the generated shares (e.g., A1, A2…A5), generating sig(s i To prevent the shares from being tampered with. The shares are... i With signature sig(s i The data is concatenated and encrypted using the AES-256 symmetric encryption algorithm (the key is negotiated via a secure channel). The encrypted share is then sent to the corresponding bank via HTTPS to ensure that the data is not stolen or tampered with during transmission. The encryption algorithm and transmission protocol described above are merely examples; other equivalent encryption and secure communication mechanisms can be used in practical applications.

[0087] After receiving and decrypting all shares from other banks, each bank aggregates them according to its own identity identifier x to form a local data share set. For example: The share set of bank A (x=1) is: {A1=87, B1=128, C1=157, D1=89, E1=115}.

[0088] The share set of bank B (x=2) is: {A2=174, B2=236, C2=224, D2=158, E2=160}.

[0089] Banks C through E follow a similar pattern.

[0090] Step 3: Local Homomorphic Cooperative Computation Within GF(503), each bank performs an additive homomorphic computation (i.e., summation) on its set of local data shares to generate a result share Sum.

[0091] Bank A calculates: SumA = (87 + 128 + 157 + 89 + 115) mod 503 = 73.

[0092] Bank B calculates: SumB = (174 + 236 + 224 + 158 + 160) mod 503 = 449.

[0093] Bank C calculates: SumC=(311+404+321+267+225) mod 503=19.

[0094] Bank D calculation: SumD=(498+129+448+416+310) mod 503=292.

[0095] Bank E calculation: SumE=(232+417+102+102+415) mod 503=262.

[0096] Each bank will store the locally computed Sum_i encrypted data and make it available only during collaboration.

[0097] Step 4: Results Summary and Reconstruction (Reconstructing Total Deposits by Using Shares from 3 Banks) Suppose bank A needs to know the total deposits. It initiates a total deposit calculation request, sends cooperation requests to B, C, D, and E, and collects SumB=449 and SumC=19. Together with its own SumA=73, it obtains a total of k=3 calculation result shares.

[0098] Bank A, within GF(503), uses points (1,73), (2,449), and (3,19) as interpolation points, and reconstructs the summation polynomial F using the Lagrange interpolation method. sum (x) and calculate its value at x=0: After calculation (division in a finite field is achieved by multiplying by the modular inverse), F is obtained. sum (0) = 400. In other implementations, any participant or designated aggregation node may act as the requester.

[0099] The reconstructed result of 4 million yuan is mathematically identical to the sum of the original deposits of the five banks (50+80+120+60+90=4 million yuan), verifying that this method can correctly complete collaborative calculations while protecting the data privacy of all parties (each bank only knows its own deposits and shares).

[0100] This embodiment clearly demonstrates the complete process of the present invention from data segmentation, secure exchange, local homomorphic computation to result reconstruction, reflecting its core characteristics and feasibility of being "usable but invisible".

[0101] The above embodiments are only used to clearly illustrate the technical solutions of this disclosure, and are not intended to limit it. Under the principles of this disclosure, corresponding adjustments and changes can be made for different application scenarios (such as joint risk control, medical research, etc.) and computational tasks (such as calculating averages, variances, etc.). These adjustments and changes should all be considered within the protection scope of this disclosure.

[0102] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0103] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.

Claims

1. A multi-party data collaborative computing method, comprising: For each participant's own original data, a constant term is constructed within a finite field, which is (k) of the original data. 1) A polynomial of degree n is used to generate n data shares, where n is the number of participants. The data shares are exchanged so that each participant obtains a set of data shares from all participants that correspond to the same value of the independent variable; Based on the data share set, homomorphic collaborative computation is performed within the finite domain to generate computation result shares; Collect k computation result shares, and calculate the shares within the finite domain using an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; wherein the final result is consistent with the result of directly performing the homomorphic collaborative computation on each original data.

2. The multi-party data collaborative computing method according to claim 1, wherein, The (k) 1) The coefficients of the polynomial, excluding the constant term, are generated by random selection within the finite field, so that fewer than k data shares cannot be used to deduce the corresponding original data.

3. The multi-party data collaborative computing method according to claim 2, wherein, The modulus of the finite field is greater than the sum of the original data of all participants.

4. The multi-party data collaborative computing method according to claim 1, wherein, The exchange of the data shares includes: The data share is encrypted using an encryption algorithm; The encrypted data shares are transmitted to the corresponding participants through a secure network transmission protocol.

5. The multi-party data collaborative computing method according to claim 4, wherein, Before performing homomorphic collaborative computation within the finite domain based on the data share set, the process includes: After receiving the encrypted data share, decrypt it.

6. The multi-party data collaborative computing method according to claim 1, wherein, The homomorphic collaborative computation includes: Without restoring the original data, perform an addition operation on the set of data shares.

7. The multi-party data collaborative computing method according to any one of claims 1-6, wherein, The method further includes: Based on another set of k different calculation result shares, the shares are calculated in the finite domain by an interpolation-based reconstruction method to obtain the verification result; Determine whether the verification result is consistent with the final result. If they are consistent, then the final result is confirmed to be correct.

8. An electronic device, comprising: One or more processors; A memory having stored one or more programs that, when executed by one or more processors, cause the one or more processors to implement the multi-party data collaborative computing method according to any one of claims 1 to 7.

9. A computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the multi-party data collaborative computing method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program or instructions that, when executed by a processor, implement the multi-party data collaborative computing method according to any one of claims 1 to 7.

11. A participant device for multi-party collaborative data computation, comprising a processor and a memory, wherein, The processor is configured to execute a program stored in the memory, enabling the participating device to: For the original data it holds, construct a constant term within a finite field that is (k) of the original data. 1) A polynomial of degree n is generated, and the function values ​​of the polynomial at n different independent variable values ​​are calculated to generate n data shares, where n is the number of participants; The data shares are exchanged with other participating devices so that each participating device obtains a set of data shares corresponding to the same independent variable value from all participating devices; Based on the data share set, homomorphic collaborative computation is performed within the finite domain to generate computation result shares; As the requesting party, it obtains the computation result shares from the devices of other participating parties, and when it obtains k computation result shares, it calculates the computation result shares in the finite domain through an interpolation-based reconstruction method to obtain the final result of the collaborative analysis; And as a respondent, in response to a request from another participating device, it provides the requesting party with its own share of the computational results.

Images