Unlock instant, AI-driven research and patent intelligence for your innovation.
Federation learning incentive method and system based on federation chain
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
An alliance and federation technology, applied in the blockchain field, can solve problems such as huge costs, achieve the effects of reducing costs, ensuring data security, and reducing blindness
Inactive Publication Date: 2020-02-21
INSPUR ARTIFICIAL INTELLIGENCE RES INST CO LTD SHANDONG CHINA
View PDF4 Cites 15 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
In reality, it is almost impossible to integrate the data scattered in various places and institutions, or the cost required is huge
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0072] combined with Figure 1-4 , this embodiment proposes a federated learning incentive method based on alliance chain, the implementation process of this method includes:
[0073] S10. Construct the transaction subject and the operation subject into an alliance chain.
[0074] In this step, the trading subjects and operating subjects involved are all nodes on the alliance chain. The main body of the transaction is the data provider and the data consumer, and the main body of the operation is a third-party organization.
[0075] In this step, the transaction subject and the operation subject are constructed into a consortium chain. This process is actually:
[0076] Data providers, consumers and third-party institutions that will participate in the intermediate data of the joint training machine learning model will build an alliance chain;
[0077] S20. Encrypt and align the user groups of the transaction subjects on the alliance chain, and determine the shared user grou...
Embodiment 2
[0099] combined with Figure 5 , this embodiment proposes a federated learning incentive system based on alliance chain, which includes:
[0100] Construction module 1, used to construct a consortium chain of transaction subjects and operation subjects, the transaction subject is a data provider and data consumer, and the operation subject is a third-party organization;
[0101] Encrypted alignment module 2, which is used to encrypt and align the user groups of transaction subjects on the alliance chain, and determine the shared user groups and shared feature dimensions;
[0102] Training convergence module 3, assisting the operating entity to train the machine learning model by using the determined common user groups and common feature dimensions, and complete the model training until the loss function converges;
[0103] Scoring incentive module 4, assisting the operating entity to score the behaviors that occur on the alliance chain, and mapping the credit points to the tr...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
PUM
Login to View More
Abstract
The invention discloses a federation learning incentive method based on a federation chain, and relates to the technical field of block chains, and the method comprises the steps: building a transaction main body and an operation main body into the federation chain; encrypting and aligning the user groups of the transaction subjects on the alliance chain, and determining a common user group and acommon feature dimension; enabling the operation main body to use the determined common user group and the common feature dimension to train a machine learning model until a loss function converges, and completing model training; and enabling the operation main body to carry out reputation scoring on behaviors occurring on the alliance chain, map reputation points to transaction costs, and stimulate each transaction main body to carry out account book maintenance through the transaction costs. The invention further discloses a federated learning excitation system based on the alliance chain, the federated learning excitation system is the same as the method, the data island problem can be solved, and the safety in the data interaction process can be guaranteed.
Description
technical field [0001] The present invention relates to the technical field of block chain, specifically a federated learning incentive method and system based on alliance chain. Background technique [0002] The consortium chain is only for members of a specific group and limited third parties, and internally designates multiple pre-selected nodes as bookkeepers. The generation of each block is jointly decided by all pre-selected nodes, and other access nodes can participate in transactions. But without asking about the bookkeeping process, other third parties can make limited inquiries through the open API of the blockchain. In order to achieve better performance, the consortium chain has certain requirements for the configuration of consensus or verification nodes and the network environment. With the admission mechanism, it is easier to improve transaction performance and avoid some problems caused by uneven participants. [0003] Federated machine learning is also kno...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q40/04H04L29/06
CPCG06Q40/04H04L63/0428
Inventor 尹青山李锐于治楼
Owner INSPUR ARTIFICIAL INTELLIGENCE RES INST CO LTD SHANDONG CHINA
Login to View More 
Login to View More