Privacy-enhanced federal decision-making method, apparatus, system, and storage medium

Abstract

The present invention provides a privacy-enhanced federated decision-making method, device, system, and storage medium, and proposes a method of training a global decision-making model under the condition of ensuring the data privacy of data terminals, and each federated data terminal reports a local decision to a federated coordinator Model information. The federated coordinator uses the local decision-making model information reported by the federated data terminals to train the global decision-making model. The trained global decision-making model can coordinate the decisions between federated data terminals, such as coordinating the decision-making sequence of federated data terminals. Or coordinate whether the federal data terminal needs to be added to the decision-making task, which not only solves the difficult problem of cross-data terminal coordination, but also improves the accuracy of the internal decision-making of the data terminal. The data terminal has an adaptive federation decision-making model, which greatly improves This increases the flexibility of federal data terminal decision-making.

Description

technical field [0001] The invention relates to the technical field of cross-data terminal decision-making, in particular to a privacy-enhanced federated decision-making method, device, system and storage medium. Background technique [0002] With the rapid development and promotion of artificial intelligence technologies such as machine learning, model training based on a large amount of raw data to obtain a model with a higher degree of fitting has become the pursuit of more enterprises and scholars. Emphasis on data privacy and security has become a worldwide trend, and all countries are strengthening the protection of data security and privacy. The new law "General Data Protection Regulation (GDPR)" introduced by the European Union clearly stipulates that any collection, The institutions and organizations that transmit, retain or process personal information related to all EU member states are bound; my country also clearly states in many laws and regulations that network...

AI Technical Summary

Problems solved by technology

In view of the fact that the internal data privacy of data terminals makes the decision-making coordination process between data terminals difficult, the local decision-making model information of each data terminal is used to train a global decision-making model to coordinate the decision-making between data terminals; for data terminals that are difficult to make accurate decisions The problem is to adopt a federated learning design architecture to train a federated decision-making model, and each data terminal can decide how to use the federated decision-making model according to the local demand for federated coordinator guidance

It not only solves the difficult problem of cross-data terminal coordination, but also improves the accuracy of the internal decision-making of the data terminal. The data terminal has an adaptive federal decision-making model, which greatly improves the flexibility of the data terminal decision-making.

Images