Trusted federal gradient boosting decision tree training method based on trusted incentive

Abstract

The invention discloses a trustworthy federated gradient boosting decision tree training method based on trustworthy incentive. The method comprises a task initiation stage, a transverse federated learning-based gradient boosting decision tree stage, a contribution value evaluation stage, a verification stage and a reward distribution stage. The invention provides a shapley value approximate calculation method suitable for the scheme, the calculation amount of the shapley value is reduced through a dynamic grouping method, the contribution value is verified by utilizing a trusted execution environment (TEE), and the fairness of contribution value calculation is ensured to a certain extent.

Description

technical field [0001] The invention relates to the technical field of Federated Learning, in particular to a credible federated gradient boosting decision tree training method based on credible incentives. Background technique [0002] In most industries, data exists in the form of isolated islands. Due to issues such as industry competition, privacy security, and complex administrative procedures, even data integration between different departments of the same company faces many obstacles. It is almost impossible to integrate data scattered in various places and institutions, or the cost required is huge. Federated learning is a machine learning framework proposed in recent years, which can effectively help multiple institutions conduct data usage and machine learning modeling while meeting the requirements of user privacy protection, data security and government regulations. Different from the traditional centralized training method, it assigns training tasks to each dat...

AI Technical Summary

Problems solved by technology

[0004] What the present invention aims to solve is the problem that the client is not willing to contribute its own data to participate in model training in the actual federated GBDT training and cannot guarantee that the client participating in the federated GBDT training task will perform correct model training. Federated Gradient Boosting Decision Tree Training Method

Images