Federated Learning for Privacy-Preserving Medical AI
arXiv:2603.15901v1 Announce Type: new Abstract: This dissertation investigates privacy-preserving federated learning for Alzheimer's disease classification using three-dimensional MRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Existing methodologies often suffer from unrealistic data partitioning, inadequate privacy guarantees, and insufficient benchmarking, limiting their practical deployment in healthcare. To address these gaps, this research proposes a novel site-aware data partitioning strategy that preserves institutional boundaries, reflecting real-world multi-institutional collaborations and data heterogeneity. Furthermore, an Adaptive Local Differential Privacy (ALDP) mechanism is introduced, dynamically adjusting privacy parameters based on training progression and parameter characteristics, thereby significantly improving the privacy-utility trade-off over traditional fixed-noise approaches. Systematic empirical evaluation across multiple client federa
arXiv:2603.15901v1 Announce Type: new Abstract: This dissertation investigates privacy-preserving federated learning for Alzheimer's disease classification using three-dimensional MRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Existing methodologies often suffer from unrealistic data partitioning, inadequate privacy guarantees, and insufficient benchmarking, limiting their practical deployment in healthcare. To address these gaps, this research proposes a novel site-aware data partitioning strategy that preserves institutional boundaries, reflecting real-world multi-institutional collaborations and data heterogeneity. Furthermore, an Adaptive Local Differential Privacy (ALDP) mechanism is introduced, dynamically adjusting privacy parameters based on training progression and parameter characteristics, thereby significantly improving the privacy-utility trade-off over traditional fixed-noise approaches. Systematic empirical evaluation across multiple client federations and privacy budgets demonstrated that advanced federated optimisation algorithms, particularly FedProx, could equal or surpass centralised training performance while ensuring rigorous privacy protection. Notably, ALDP achieved up to 80.4% accuracy in a two-client configuration, surpassing fixed-noise Local DP by 5-7 percentage points and demonstrating substantially greater training stability. The comprehensive ablation studies and benchmarking establish quantitative standards for privacy-preserving collaborative medical AI, providing practical guidelines for real-world deployment. This work thereby advances the state-of-the-art in federated learning for medical imaging, establishing both methodological foundations and empirical evidence necessary for future privacy-compliant AI adoption in healthcare.
Executive Summary
This dissertation proposes a novel federated learning approach for Alzheimer's disease classification using 3D MRI data. A site-aware data partitioning strategy and Adaptive Local Differential Privacy (ALDP) mechanism are introduced to address existing limitations in privacy-preserving federated learning. The study demonstrates that ALDP achieves higher accuracy and training stability compared to traditional fixed-noise approaches. The findings establish quantitative standards for privacy-preserving collaborative medical AI, providing practical guidelines for real-world deployment. The research advances the state-of-the-art in federated learning for medical imaging, laying the groundwork for future privacy-compliant AI adoption in healthcare. The study's results have significant implications for the development of trustworthy medical AI systems.
Key Points
- ▸ Introduction of site-aware data partitioning strategy to preserve institutional boundaries
- ▸ Adaptive Local Differential Privacy (ALDP) mechanism for dynamic privacy parameter adjustment
- ▸ Systematic empirical evaluation of federated optimisation algorithms and privacy protection
Merits
Methodological Innovation
The dissertation introduces novel site-aware data partitioning and ALDP mechanisms, addressing existing gaps in privacy-preserving federated learning.
Demerits
Limited Generalizability
The study focuses on Alzheimer's disease classification using 3D MRI data, which may limit the generalizability of the findings to other medical imaging applications.
Expert Commentary
This dissertation makes a significant contribution to the field of federated learning for medical imaging. The introduction of site-aware data partitioning and ALDP mechanisms addresses critical limitations in existing approaches, enabling the development of more robust and trustworthy medical AI systems. The study's findings have far-reaching implications for the healthcare sector, where the adoption of AI systems is increasingly dependent on ensuring rigorous privacy protection. As the field continues to evolve, it is essential to address the challenges associated with data protection and privacy, and this research provides a crucial step towards that goal.
Recommendations
- ✓ Future studies should investigate the generalizability of the proposed approach to other medical imaging applications.
- ✓ Research institutions and healthcare organizations should collaborate to develop and deploy privacy-preserving federated learning frameworks for medical AI systems.