Inclusive, Differentially Private Federated Learning for Clinical Data

2025-05-28

AI Summary

To address the imbalanced model performance in medical federated learning—arising from the coexistence of privacy preservation requirements, resource heterogeneity across institutions, and regulatory divergence—this paper proposes a compliance-aware differentially private federated learning framework. Methodologically, it introduces three core innovations: (1) a quantifiable client-side compliance scoring system grounded in established medical data security standards; (2) a noise-adaptive injection mechanism enabling dynamic, demand-driven allocation of privacy budgets; and (3) an inclusive participation paradigm that enhances the effective contribution of low-resource institutions. Experimental evaluation on multi-center clinical datasets demonstrates that, compared to conventional approaches, the proposed framework achieves up to a 15% improvement in model accuracy under collaborative training involving both highly regulated and resource-constrained institutions. Moreover, it significantly enhances model generalizability and deployment feasibility in real-world healthcare settings.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Data transparency and provenance
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

Federated Learning (FL) offers a promising approach for training clinical AI models without centralizing sensitive patient data. However, its real-world adoption is hindered by challenges related to privacy, resource constraints, and compliance. Existing Differential Privacy (DP) approaches often apply uniform noise, which disproportionately degrades model performance, even among well-compliant institutions. In this work, we propose a novel compliance-aware FL framework that enhances DP by adaptively adjusting noise based on quantifiable client compliance scores. Additionally, we introduce a compliance scoring tool based on key healthcare and security standards to promote secure, inclusive, and equitable participation across diverse clinical settings. Extensive experiments on public datasets demonstrate that integrating under-resourced, less compliant clinics with highly regulated institutions yields accuracy improvements of up to 15% over traditional FL. This work advances FL by balancing privacy, compliance, and performance, making it a viable solution for real-world clinical workflows in global healthcare.

Problem

• Enhance privacy in federated learning for clinical data
• Adaptively adjust noise based on client compliance scores
• Improve accuracy and inclusivity in diverse clinical settings

Innovation

• Compliance-aware FL framework with adaptive DP noise
• Compliance scoring tool based on healthcare standards
• Integrates diverse clinics for improved accuracy