Browse Teach Curate

Privacy-Adaptive Federated Learning with Blockchain Provenance for Internet of Vehicles

FAPL-DM-BC integrates adaptive privacy-aware federated learning and dynamic masking (tuning the privacy mechanism to real-time data sensitivity), blockchain smart-contract provenance and decentralized validation against model poisoning, secure aggregation via FedAvg and secure multi-party computation, and an explainable-AI feedback loop, into one architecture for secure and interpretable learning across vehicular networks.

Boomer Text

FAPL-DM-BC: A Secure and Scalable FL Framework with Adaptive Privacy and Dynamic Masking

Proposes FAPL-DM-BC, an integrated federated-learning framework for the Internet of Vehicles (IoV) that combines four ideas: Federated Adaptive Privacy-Aware Learning with Dynamic Masking, which adju…