4.7 Article

When Federated Learning Meets Blockchain: A New Distributed Learning Paradigm

Journal

IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE
Volume 17, Issue 3, Pages 26-33

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/MCI.2022.3180932

Keywords

-

Funding

  1. National Natural Science Foundation of China [62002170, 61872184]
  2. Fundamental Research Funds for the Central Universities [30919011274]
  3. Natural Science Foundation of Jiangsu Province [BK20210331]
  4. Jiangsu Specially-Appointed Professor Program in 2021
  5. Natural Science Fund of Guangdong Province [2020A1515010708]
  6. Natural Science Fund of Shenzhen [JCYJ20210324094609027]
  7. U.S. National Science Foundation [ECCS-2039716, CNS-2107216, CNS-2128368]

Ask authors/readers for more resources

Motivated by end-user computing capabilities and privacy concerns, a blockchain-assisted decentralized FL framework is proposed to address the single point of failure in conventional FL frameworks and prevent malicious clients. The framework integrates model aggregation and blockchain mining tasks, and explores privacy and resource allocation issues. A critical issue, lazy clients, is also identified and solutions involving local differential privacy and pseudo-noise sequences are discussed.
Motivated by the increasingly powerful computing capabilities of end-user equipment, and by the growing privacy concerns over sharing sensitive raw data, a distributed machine learning paradigm known as federated learning (FL) has emerged. By training models locally at each client and aggregating learning models at a central server, FL has the capability to avoid sharing data directly, thereby reducing privacy leakage. However, the conventional FL framework relies heavily on a single central server, and it may fail if such a server behaves maliciously. To address this single point of failure, in this work, a blockchain-assisted decentralized FL framework is investigated, which can prevent malicious clients from poisoning the learning process, and thus provides a self-motivated and reliable learning environment for clients. In this framework, the model aggregation process is fully decentralized and the tasks of training for FL and mining for blockchain are integrated into each participant. Privacy and resource-allocation issues are further investigated in the proposed framework, and a critical and unique issue inherent in the proposed framework is disclosed. In particular, a lazy client can simply duplicate models shared by other clients to reap benefits without contributing its resources to FL. To address these issues, analytical and experimental results are provided to shed light on possible solutions, i.e., adding noise to achieve local differential privacy and using pseudo-noise (PN) sequences as watermarks to detect lazy clients.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available