Staleness-Bounded Aggregation for Cross-Institutional Federated Clinical Models
Main Article Content
Abstract
Federated learning can also enable development of cross-institutional clinical prediction models which do not require sharing of sensitive information about patients, which can help to improve privacy and regulatory quality. However, the cross institutional deployments are often known to have dissimilar computational attributes, network delays, and imperfect involvement of customers thereby receiving stale model refreshments which are harmful to convergence and predictive speed. The given paper suggests a new Staleness-Bounded Aggregation (SBA) system, where the influence of stalia news that the clients gather with the assistance of the dynamic staleness threshold and weighting according to the recency in the aggregation process on the server is limited. The proposed solution is efficient and balances model freshness and efficiency of communication, though it keeps the advantages of privacy of federated learning. The framework is experimented using non-identical (non-IID) simulated multi-hospital settings with clinical data having varying delay on communications and client availability. The experimental results indicate that converging SBA is quicker, more exact of categorization and more robust than the conventional Federated Averaging and other existing asynchronous aggregation frameworks. Specifically, the specified technique will reduce the adverse effect of slow updates, boost the stability of the most diverse environments, and reduce communication expenses without increasing the privacy concern. These results suggest that staleness-bounded aggregation is a realistic and scalable method of addressing distributed clinical machine learning, which is viable to support cross-institutional cooperation, and will not affect the quality of the predictors of the actual healthcare system. The suggested organization is likely to rely on credible, robust, and safe-next-generation healthcare analytics federated clinical intelligences.
Article Details
Section
How to Cite
References
[1] H. B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, "Communication-Efficient Learning of Deep Networks from Decentralized Data," in Proc. 20th Int. Conf. Artificial Intelligence and Statistics (AISTATS), Fort Lauderdale, FL, USA, 2017, pp. 1273–1282.
[2] B. Liu et al., "Recent Advances on Federated Learning: A Systematic Survey," Neurocomputing, vol. 585, pp. 127695, 2024.
[3] R. Lu et al., "Adaptive Asynchronous Federated Learning," Future Generation Computer Systems, vol. 155, pp. 313–325, 2024.
[4] M. Chen et al., "FedSA: A Staleness-Aware Asynchronous Federated Learning Algorithm with Non-IID Data,"Future Generation Computer Systems, vol. 120, pp. 1–12, 2021.
[5] J. Liu, J. Jia, T. Che, C. Huo, J. Ren, Y. Zhou, H. Dai, and D. Dou, "FedASMU: Efficient Asynchronous Federated Learning with Dynamic Staleness-Aware Model Update," Proc. AAAI Conf. Artificial Intelligence, vol. 38, no. 12, pp. 13900–13908, 2024.
[6] S. Sun et al., "Staleness-Controlled Asynchronous Federated Learning: Accuracy and Efficiency Tradeoff," IEEE Transactions on Mobile Computing, vol. 23, no. 9, pp. 9023–9038, 2024.
[7] Y. Fraboni, R. Vidal, L. Kameni, and M. Lorenzi, "A General Theory for Federated Optimization with Asynchronous and Heterogeneous Clients Updates," Journal of Machine Learning Research, vol. 24, pp. 1–54, 2023.
[8] Z. Chen et al., "Adaptive Semi-Asynchronous Federated Learning over Wireless Networks," IEEE Transactions on Communications, vol. 72, no. 8, pp. 5103–5118, 2024
[9] C.-H. Hu et al., "Scheduling and Aggregation Design for Asynchronous Federated Learning over Wireless Networks," IEEE Journal on Selected Areas in Communications, vol. 41, no. 7, pp. 2203–2218, 2023.
[10] Q. Ma et al., "FedSA: A Semi-Asynchronous Federated Learning Mechanism in Heterogeneous Edge Computing," IEEE Journal on Selected Areas in Communications, vol. 39, no. 12, pp. 3654–3667, 2021.
[11] Y. Zhang et al., "FedMDS: An Efficient Model Discrepancy-Aware Semi-Asynchronous Clustered Federated Learning Framework," IEEE Transactions on Parallel and Distributed Systems, vol. 34, no. 11, pp. 3236–3250, 2023.
[12] Y. Li et al., "Convergence Analysis of Sequential Federated Learning on Heterogeneous Data," in Advances in Neural Information Processing Systems (NeurIPS), vol. 37, 2024.
[13] W. Huang et al., "Federated Learning for Generalization, Robustness, Fairness: A Survey and Benchmark," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 10, pp. 6789–6810, 2024.
[14] M. M. Islam et al., "Multi-Level Feature Fusion for Multimodal Human Activity Recognition in Internet of Healthcare Things," Information Fusion, vol. 93, pp. 102012, 2023.