A financial data supervision method based on privacy computing
By fusing the front-back propagation base model and the gradient-enhanced base model into a meta-model, and using a shared algorithm, the problem that traditional financial data supervision methods are difficult to balance privacy protection and data utilization is solved, and efficient and secure financial data supervision is achieved.
CN118070307BActive Publication Date: 2025-04-04CHENGDU UNIV OF INFORMATION TECH +1
2 Cites 0 Cited by
Patent Information
- Application Number
- CN202410189385.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-02-20
- Publication Date
- 2025-04-04
- Estimated Expiration
- 2044-02-20
AI Technical Summary
Technical Problem
Traditional financial data supervision methods are difficult to balance privacy protection with effective data utilization, and cannot effectively deal with rapidly changing financial markets and data security threats.
Method used
The financial data supervision method based on privacy computing is adopted, and the forward-backward propagation base model and gradient-enhanced base model are fused into meta-models, and the sharing algorithm is used to realize the supervision of financial data.
Benefits of technology
It realizes sharing and analyzing financial data without leaking sensitive information, reducing the risk of data leakage, improving the availability and security of data, and enhancing the stability and efficiency of financial data supervision.
✦ Generated by Eureka AI based on patent content.
Abstract
The present invention discloses a financial data supervision method based on privacy computing, which includes: preprocessing the financial data information tables of different regulatory agency nodes and encrypting the financial data; building a forward and backward propagation base model for financial data supervision using the preprocessed data of the child nodes; building a gradient enhancement base model for financial data supervision using the preprocessed data of the child nodes; fusing the financial data supervision gradient enhancement model and the forward and backward propagation model to establish a financial data supervision meta-model, and uploading the parameter vector sequence of the financial data supervision meta-model of the child nodes; the central node aggregates the parameter vector sequence of the financial data supervision meta-model uploaded by the child nodes, updates the financial data supervision large model of the central node and shares the updated financial data supervision large model with the child nodes.
Need to check novelty before this filing date? Find Prior Art