CN113762528A - Block chain-based federal credit assessment method - Google Patents
Block chain-based federal credit assessment method Download PDFInfo
- Publication number
- CN113762528A CN113762528A CN202111059225.0A CN202111059225A CN113762528A CN 113762528 A CN113762528 A CN 113762528A CN 202111059225 A CN202111059225 A CN 202111059225A CN 113762528 A CN113762528 A CN 113762528A
- Authority
- CN
- China
- Prior art keywords
- block chain
- model
- gradient
- participating
- assessment method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000004364 calculation method Methods 0.000 claims abstract description 14
- 230000002776 aggregation Effects 0.000 claims abstract description 7
- 238000004220 aggregation Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 2
- 230000004931 aggregating effect Effects 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Bioethics (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a block chain-based federal credit assessment method, which specifically comprises the following steps: s01, each participating calculator registers and authenticates on the blockchain, and issues certificates; s02, each participating calculator downloads an initial model and a program from the block chain, and performs self-initialization self-parameter and encrypted sample ID alignment operation; s03, each participating calculation party performs gradient calculation locally, and encrypts and sends the calculation gradient result to the block chain; s04, the consensus node verifies the encryption gradient, and after consensus is achieved, the encryption gradient data are recorded in a newly generated block; s05, the intelligent contract aggregates the model parameters, aggregates the model parameters and updates the whole model; after the intelligent convergence completes the aggregation of each round of parameters, calculating the contribution value and the credit value of each participating calculating party, and recording the result on a block chain; and S06, judging whether the preset convergence condition of the model is reached by the intelligent contract.
Description
Technical Field
The invention relates to the field of block chains, in particular to a block chain-based federal credit assessment method.
Background
The blockchain ensures that the uplink data is difficult to be tampered, and the data is stored in each node, so that the node votes to maintain consistency. Smart contracts are executable code, which is an execution job, with data security provided by the blockchain.
The federated learning system is used as a learning method for a plurality of clients to contribute local data to cooperatively train a unified model, the data does not move, the problem that credit assessment data is limited and shared and the privacy of the data is easily violated can be solved, and the federated learning system is limited in that the model only depends on a single central server and is easily influenced by server faults. The purpose of traditional federal learning is to obtain a globally shared model for all participants. However, when the data distribution of each participant is inconsistent, the global model cannot meet the performance requirement of each federal learning participant, and some participants cannot even obtain a model better than a model trained by only using local data.
Based on the above, a block chain-based federal credit assessment method is provided, which solves the problems of limited credit data and improved privacy protection of the data.
Disclosure of Invention
In view of the above-mentioned defects in the prior art, the technical problem to be solved by the present invention is to provide a block chain-based federal credit assessment method aiming at the needs and disadvantages of the current technical development.
Firstly, the invention provides a block chain-based federal credit assessment method, which adopts the following technical scheme for solving the technical problems:
s01, registering each participating calculator on the blockchain, authenticating each calculator by the blockchain, and issuing a certificate;
s02, each participating calculator downloads an initial model and a program from the block chain, and performs self-initialization self-parameter and encrypted sample ID alignment operation;
s03, each participating calculation party performs gradient calculation locally, and encrypts and sends the calculation gradient result to the block chain;
s04, the consensus node verifies the encryption gradient, and after consensus is achieved, the encryption gradient data are recorded in a newly generated block;
s05, the intelligent contract aggregates the model parameters, aggregates the model parameters and updates the whole model; after the intelligent convergence completes the aggregation of each round of parameters, calculating the contribution value and the credit value of each participating calculating party, and recording the result on a block chain;
and S06, judging whether the preset convergence condition of the model is reached by the intelligent contract, if not, carrying out the next round of training, and if so, terminating the training.
Specifically, the blockchain system is a federation chain or a private chain, and its main functions are:
(1) authenticating each participant and issuing a certificate;
(2) recording encryption gradient results uploaded by each participant;
(3) recording model parameters;
(4) the contribution values and reputation values of the participants are recorded.
Specifically, the main functions of the intelligent contract system are as follows:
(1) the intelligent contract aggregates the uploaded encryption gradient and updates the model parameters;
(2) aggregating the model parameters, and updating the whole model;
(3) calculating the contribution value and the reputation value of each participant;
(4) judging whether a preset convergence condition of the model is reached, if not, carrying out next round of training, and if so, terminating the training; if not, the next round of training is performed.
The party participating in the calculation in step S01 may be a credit institution, a financial institution, a government institution, a commercial company, or the like. The step S02 encrypts the sample ID alignment in a manner that may be Freedman protocol or Diffie-Hellman key exchange protocol or other manners. The encryption and aggregation of the gradient obtained by each round of training are calculated by using safe multiple parties. The consensus algorithm in step S04 uses a byzantine-type consensus algorithm. The step S05 is to aggregate the received model parameters according to the record on the blockchain by the intelligent contract, and update the entire model. The received model parameters are updated, either using weighted averaging or median values. The contribution value and the reputation value are calculated by gradient contribution and loyalty to take into account participant fairness and efficiency. And the preset convergence condition of the model in the step S06 is determined according to parameters such as the set maximum iteration number, the loss value and the like during task training.
Drawings
Some specific embodiments of the invention will be described in detail hereinafter, by way of illustration and not limitation, with reference to the accompanying drawings. The same reference numbers in the drawings identify the same or similar elements or components. Those skilled in the art will appreciate that the drawings are not necessarily drawn to scale. The objects and features of the present invention will become more apparent in view of the following description taken in conjunction with the accompanying drawings, in which:
fig. 1 is a flow chart of a block chain-based federal credit assessment method according to the present invention.
Detailed Description
In order to clearly illustrate the present invention and make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, so that those skilled in the art can implement the technical solutions in reference to the description text. The technology of the present invention will be described in detail below with reference to the accompanying drawings in conjunction with specific embodiments.
The embodiment of the invention provides a block chain-based federal credit assessment method, which comprises the following implementation processes:
and S01, the participating calculators C1, C2, C3 and C4 register to the blockchain B respectively, the blockchain B verifies the registration information of each participating calculator, and after the registration information is authenticated, certificates are issued to each participating calculator.
S02, the user U1 creates an intelligent contract S1 according to the intelligent contract template of credit evaluation, the intelligent contract S1 runs a block chain B, a credit evaluation task T is created, and a training model M and parameters w are initialized.
Because credit assessment is to longitudinally aggregate multiple data characteristics for the same user set, the method belongs to a longitudinal federal learning method, besides the sample IDs shared by the participants, other sample IDs cannot be revealed to each other, the process needs to encrypt the sample IDs to be aligned, the encryption mode can be a Freedman protocol or a Diffie-Hellman key exchange protocol or other modes, the Freedman protocol is adopted, the core is to encrypt the polynomial coefficients by using addition homomorphic encryption, and only the sample IDs in the intersection can enable the polynomial to take the value of zero. The Diffie-Hellman key exchange protocol requires finding an encryption algorithm that satisfies the requirement that two successive encryption operations can exchange the order.
S03, each participating calculator C1, C2, C3 and C4 carries out gradient calculation locally, and the calculation gradient result is encrypted and sent to a block chain;
gradient encryption and aggregation employ secure multiparty computation, each participant CiSecret input gradient thetai，The common execution function F is (theta)1,θ2,θ3,θ4)→(δ1,δ2,δ3,δ4) In which the gradient delta is encryptediIs participant CiThe corresponding output obtained is that any participant cannot know any input information of other participants in the calculation process of the function F.
S04, the consensus node verifies the encryption gradient, and after consensus is achieved, the encryption gradient data are recorded in a newly generated block;
s05, the intelligent contract aggregates the model parameters, aggregates the model parameters and updates the whole model; and after each round of parameter aggregation is completed, calculating the contribution value and the credit degree of each participating calculating party, and recording the result on the block chain.
The contribution value c is determined according to the contribution of the participant to the gradient, and the calculation of the reputation t is determined according to the loyalty of the participant to comply with the system protocol.
In the next round, only the participants with higher contribution values and credibility have the opportunity to become the consensus nodes, so that all the participants are stimulated to obey the system protocol.
And S06, judging whether a preset convergence condition of the model is reached by the intelligent contract, wherein the convergence condition is determined according to parameters such as the set maximum iteration times, the loss value and the like during task training. If not, the next round of training is performed, and if so, the training is terminated.
Claims (9)
1. A block chain-based federal credit assessment method is characterized in that the implementation process of the method comprises the following steps:
s01, registering each participating calculator on the blockchain, authenticating each calculator by the blockchain, and issuing a certificate;
s02, each participating calculator downloads an initial model and a program from the block chain, and performs self-initialization self-parameter and encrypted sample ID alignment operation;
s03, each participating calculation party performs gradient calculation locally, and encrypts and sends the calculation gradient result to the block chain;
s04, the consensus node verifies the encryption gradient, and after consensus is achieved, the encryption gradient data are recorded in a newly generated block;
s05, the intelligent contract aggregates the model parameters, aggregates the model parameters and updates the whole model; after the intelligent convergence completes the aggregation of each round of parameters, calculating the contribution value and the credit value of each participating calculating party, and recording the result on a block chain;
and S06, judging whether the preset convergence condition of the model is reached by the intelligent contract, if not, carrying out the next round of training, and if so, terminating the training.
2. The block chain-based federal credit assessment method as claimed in claim 1, wherein the block chain system is a federation chain or a private chain, and its main functions are:
(1) authenticating each participant and issuing a certificate;
(2) recording encryption gradient results uploaded by each participant;
(3) recording model parameters;
(4) the contribution values and reputation values of the participants are recorded.
3. A block chain-based federal credit assessment method as claimed in claim 1, wherein the main functions of the intelligent contract system are:
(1) the intelligent contract aggregates the uploaded encryption gradient and updates the model parameters;
(2) aggregating the model parameters, and updating the whole model;
(3) the contribution values and reputation values of the various parties are calculated.
(4) Judging whether a preset convergence condition of the model is reached, if not, carrying out next round of training, and if so, terminating the training; if not, the next round of training is performed.
4. The block chain-based federal credit assessment method as claimed in claim 1, wherein the participating party in step S01 is a credit agency, a financial agency, a government agency, a commercial company, or the like.
5. The block chain-based federal credit assessment method as claimed in claim 1, wherein said step S02 encrypts sample ID alignment in a manner selected from Freedman protocol, Diffie-Hellman key exchange protocol, and others.
6. The block chain-based federal credit assessment method as claimed in claim 1, wherein the gradient encryption and aggregation obtained from each training pass is performed using secure multiparty computation.
7. The method and system according to claim 1, wherein the consensus algorithm in step S04 uses a byzantine-type consensus algorithm.
8. The blockchain-based federal credit assessment method of claim 1, wherein the step S05 is implemented to aggregate the received model parameters according to the records on the blockchain by an intelligent contract, and update the entire model. The received model parameters are updated, either using weighted averaging or median values. The contribution value and the reputation value are calculated by gradient contribution and loyalty to take into account participant fairness and efficiency.
9. The block chain-based federal credit assessment method as claimed in claim 1, wherein the preset convergence condition of the model in step S06 is determined according to parameters such as the maximum iteration number and the loss value set during task training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111059225.0A CN113762528A (en) | 2021-09-10 | 2021-09-10 | Block chain-based federal credit assessment method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111059225.0A CN113762528A (en) | 2021-09-10 | 2021-09-10 | Block chain-based federal credit assessment method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113762528A true CN113762528A (en) | 2021-12-07 |
Family
ID=78794648
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111059225.0A Pending CN113762528A (en) | 2021-09-10 | 2021-09-10 | Block chain-based federal credit assessment method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113762528A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114726529A (en) * | 2022-04-06 | 2022-07-08 | 湘潭大学 | Smart power grid data aggregation method based on credit consensus mechanism |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861152A (en) * | 2021-02-08 | 2021-05-28 | 北京航空航天大学 | Federal learning incentive method and system based on permit chain |
| CN113052331A (en) * | 2021-02-19 | 2021-06-29 | 北京航空航天大学 | Block chain-based Internet of things personalized federal learning method |
| CN113095510A (en) * | 2021-04-14 | 2021-07-09 | 深圳前海微众银行股份有限公司 | Block chain-based federal learning method and device |
| WO2021159753A1 (en) * | 2020-09-21 | 2021-08-19 | 平安科技（深圳）有限公司 | Federated learning model training method and apparatus, terminal device, and storage medium |
| CN113347162A (en) * | 2021-05-19 | 2021-09-03 | 中央财经大学 | Block link point contribution degree proving consensus method for crowd-sourcing service |
-
2021
- 2021-09-10 CN CN202111059225.0A patent/CN113762528A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021159753A1 (en) * | 2020-09-21 | 2021-08-19 | 平安科技（深圳）有限公司 | Federated learning model training method and apparatus, terminal device, and storage medium |
| CN112861152A (en) * | 2021-02-08 | 2021-05-28 | 北京航空航天大学 | Federal learning incentive method and system based on permit chain |
| CN113052331A (en) * | 2021-02-19 | 2021-06-29 | 北京航空航天大学 | Block chain-based Internet of things personalized federal learning method |
| CN113095510A (en) * | 2021-04-14 | 2021-07-09 | 深圳前海微众银行股份有限公司 | Block chain-based federal learning method and device |
| CN113347162A (en) * | 2021-05-19 | 2021-09-03 | 中央财经大学 | Block link point contribution degree proving consensus method for crowd-sourcing service |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114726529A (en) * | 2022-04-06 | 2022-07-08 | 湘潭大学 | Smart power grid data aggregation method based on credit consensus mechanism |
| CN114726529B (en) * | 2022-04-06 | 2024-09-24 | 湘潭大学 | Smart grid data aggregation method based on credibility consensus mechanism |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108418783B (en) | Method and medium for protecting privacy of intelligent contracts of block chains | |
| CN107196762B (en) | Big data oriented power determining method | |
| CN112861152A (en) | Federal learning incentive method and system based on permit chain | |
| WO2020259156A1 (en) | Blockchain-based private transaction method and apparatus | |
| CN111242453B (en) | Financial evaluation method and evaluation system based on block chain | |
| CN112232527A (en) | Safe distributed federal deep learning method | |
| CN113111124B (en) | Block chain-based federal learning data auditing system and method | |
| CN112839046B (en) | Traceable anonymous crowdsourcing method and system based on block chain | |
| WO2022156594A1 (en) | Federated model training method and apparatus, electronic device, computer program product, and computer-readable storage medium | |
| CN112187712A (en) | Anonymous authentication method and system for trust in de-center mobile crowdsourcing | |
| CN114386069B (en) | Federal learning model training method based on conditional privacy set intersection | |
| Huang et al. | DAPA: A decentralized, accountable, and privacy-preserving architecture for car sharing services | |
| CN115270145A (en) | User electricity stealing behavior detection method and system based on alliance chain and federal learning | |
| CN113420886B (en) | Training method, device, equipment and storage medium for longitudinal federal learning model | |
| CN113762528A (en) | Block chain-based federal credit assessment method | |
| Keshavarzkalhori et al. | Federify: a verifiable federated learning scheme based on zksnarks and blockchain | |
| CN108259180B (en) | Method for quantum specifying verifier signature | |
| CN110865793A (en) | Method for generating random number based on block chain intelligent contract | |
| CN113806764B (en) | Distributed support vector machine based on blockchain and privacy protection and optimization method thereof | |
| Xu et al. | Efficient privacy-preserving electronic voting scheme based on blockchain | |
| CN117290887A (en) | Account blockchain-based accountability privacy protection intelligent contract implementation method | |
| CN117521102A (en) | Model training method and device based on federal learning | |
| CN114866289B (en) | Privacy credit data security protection method based on alliance chain | |
| CN113822758B (en) | Self-adaptive distributed machine learning method based on blockchain and privacy | |
| CN112054891B (en) | Block chain-based common user determination method, electronic device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211207 |
|
| WD01 | Invention patent application deemed withdrawn after publication |