Equitable Federated Learning with NCA

2025-06-26

AI Summary

Federated learning (FL) faces significant deployment challenges in low- and middle-income countries (LMICs) due to scarce medical infrastructure, resource-constrained edge devices, and unstable network connectivity—hindering equitable adoption for medical image segmentation. Method: We propose the first lightweight, low-communication, encryption-ready FL framework tailored for medical image segmentation. Our approach integrates the compact Med-NCA neural cellular automata architecture into FL, synergistically combining model compression, differential privacy–ready communication protocols, and an edge-cloud collaborative training paradigm, alongside an optimized FedAvg variant. Contribution/Results: Evaluated on multi-center medical imaging datasets, our framework achieves a model size <2 MB, reduces per-round communication volume by 73%, ensures stable convergence under 4G-grade network conditions, and supports end-to-end encrypted deployment. This markedly enhances the practicality and equitable accessibility of FL in resource-limited settings.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Segmentation
• Natural Language Processing: Safety and Robustness

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Federated Learning (FL) is enabling collaborative model training across institutions without sharing sensitive patient data. This approach is particularly valuable in low- and middle-income countries (LMICs), where access to trained medical professionals is limited. However, FL adoption in LMICs faces significant barriers, including limited high-performance computing resources and unreliable internet connectivity. To address these challenges, we introduce FedNCA, a novel FL system tailored for medical image segmentation tasks. FedNCA leverages the lightweight Med-NCA architecture, enabling training on low-cost edge devices, such as widely available smartphones, while minimizing communication costs. Additionally, our encryption-ready FedNCA proves to be suitable for compromised network communication. By overcoming infrastructural and security challenges, FedNCA paves the way for inclusive, efficient, lightweight, and encryption-ready medical imaging solutions, fostering equitable healthcare advancements in resource-constrained regions.

Problem

• Enables FL in LMICs with limited computing resources
• Addresses unreliable internet in medical image segmentation
• Provides lightweight, secure FL for healthcare in LMICs

Innovation

• Lightweight Med-NCA architecture for edge devices
• Minimizes communication costs in federated learning
• Encryption-ready for compromised network communication