A Secure Federated Learning Framework for 5G Networks
A blockchain-based secure FL framework to create smart contracts and prevent malicious or unreliable participants from being involved in FL is proposed, which can effectively deter poisoning and membership inference attacks, thereby improving the security of FL in 5G networks.
Abstract
Federated Learning (FL) has been recently proposed as an emerging paradigm to\nbuild machine learning models using distributed training datasets that are\nlocally stored and maintained on different devices in 5G networks while\nproviding privacy preservation for participants. In FL, the central aggregator\naccumulates local updates uploaded by participants to update a global model.\nHowever, there are two critical security threats: poisoning and membership\ninference attacks. These attacks may be carried out by malicious or unreliable\nparticipants, resulting in the construction failure of global models or privacy\nleakage of FL models. Therefore, it is crucial for FL to develop security means\nof defense. In this article, we propose a blockchain-based secure FL framework\nto create smart contracts and prevent malicious or unreliable participants from\ninvolving in FL. In doing so, the central aggregator recognizes malicious and\nunreliable participants by automatically executing smart contracts to defend\nagainst poisoning attacks. Further, we use local differential privacy\ntechniques to prevent membership inference attacks. Numerical results suggest\nthat the proposed framework can effectively deter poisoning and membership\ninference attacks, thereby improving the security of FL in 5G networks.\n