Blockchain-Enabled Multi-stage Incentive Framework for Federated Learning
- Publication Type:
- Thesis
- Issue Date:
- 2025
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
In federated learning scenarios where both data and model owners make significant contributions, traditional incentive models often fail to fairly assess the value of these various inputs – especially intangible efforts. In this research, we address this critical gap by introducing a novel framework that combines a multi-stage incentive mechanism, a blockchain-based clearing protocol, and a contribution buy-back method. The multi-stage incentive mechanism optimises compensation based on both quantifiable and unquantifiable contributions from data and model owners. At the same time, the blockchain-based clearing protocol facilitates trustless reward distribution, model ownership transfer, and cost-effective settlements. In addition, the framework is also compatible with various existing contribution assessment mechanisms through the contribution buyback method, mitigating risks arising from data and model incompatibility and promoting reliable collaboration. This research significantly advances federated learning by promoting fair compensation, security, and ethical practices, enabling broader adoption across various domains. For example, in healthcare, this approach can enable secure and equitable collaboration between hospitals, healthcare facilities, and machine learning experts, advancing goals like predictive analytics while ensuring data privacy and regulatory compliance.
Please use this identifier to cite or link to this item: