Strategies Module#
Federated learning strategies for contribution evaluation and compensation.
This package provides the core strategies for Rizemind’s federated learning framework. It includes mechanisms for both evaluating participant contributions and distributing rewards based on those contributions.
The package contains two main submodules:
compensation: Strategies for distributing rewards to participants based on their contributions. These integrate with blockchain-based payment systems and Flower federated learning strategies.
contribution: Tools for evaluating and quantifying how much each participant contributes to the overall model performance. This includes Shapley value calculations, sampling strategies, and both centralized and decentralized evaluation approaches.
Together, these strategies enable fair and transparent federated learning by connecting contribution measurement to appropriate compensation mechanisms.
Contents: