IEEE 3127-2025

To provide an architectural framework and application guidelines for blockchain-based federated machine learning (BC-FML), including the following:--A description and a definition of BC-FML--The types for BC-FML--Application scenarios for each type--A definition of capability for BC-FML and guidelines for evaluating these systems--Security and privacy guidelines of BC-FML--Performance evaluation of BC-FML in real application systems

The purpose of this document is to provide guidance for improving the security audibility and traceability of BC-FML.BC-FML helps data owners, coordinators, model users, etc., to realize multi-party federated modeling while meeting applicable interaction, decentralization, safety, reliability, and robustness requirements. BC-FML can improve the privacy for data owners, coordinators, model users, etc., and enable those entities to permit functions including the use of data, withdrawing the use of data, and potentially selling data under specified conditions.

New IEEE Standard - Active. Guidance for improving the security auditability and traceability of blockchain-based federated machine learning is provided in this document. Blockchain-based federated machine learning helps data owners, producers, consumers, and collaborators to realize multi-party secure computing while meeting applicable interaction, decentralization, safety, reliability, and robustness guidelines. Blockchain-based Federated Machine Learning can improve the privacy of data owners, producers, consumers, and collaborators, and enable those entities to give permission for functions including the use of data, withdrawing the use of data, and potentially selling data under specified conditions.