In recent years, federated learning has gained significant attention due to its ability to train machine learning models on distributed data without compromising data privacy. However, the traditional utility-privacy trade-off in federated learning often leads to suboptimal solutions that prioritize one aspect over the other. To address this issue, this paper proposes a new method called efficiency constrained utility-privacy bi-objective optimization in federated learning.
The proposed method considers both efficiency and privacy as equally important objectives and aims to find a balance between them. The authors propose a novel optimization algorithm that incorporates efficiency constraints to ensure the solution is efficient while maintaining privacy. They also introduce a new metric called the "efficient privacy loss" to measure the trade-off between efficiency and privacy.
The proposed method is evaluated on several real-world datasets and shows improved performance compared to existing methods. The authors demonstrate that their approach can find a better balance between efficiency and privacy, leading to more accurate model training while protecting sensitive information.
Conclusion: In this paper, the authors propose a novel optimization algorithm for federated learning that considers both efficiency and privacy as equally important objectives. By incorporating efficiency constraints and introducing a new metric called efficient privacy loss, the proposed method achieves a better balance between these two aspects. The results show that the proposed method outperforms existing methods in terms of accuracy while maintaining privacy. This work has significant implications for real-world applications where data privacy is critical, such as medical research or financial transactions.
