Energy-Efficient Federated Learning via Adaptive Encoder Freezing for MRI-to-CT Conversion: A Green AI-Guided Research

2025-11-25

AI Summary

Federated learning (FL) for medical image translation—particularly MRI-to-CT synthesis—imposes high computational and energy demands, hindering participation by resource-constrained institutions and exacerbating healthcare inequity. Method: Addressing green AI objectives, we propose the first adaptive encoder freezing strategy for federated MRI→CT translation. It dynamically monitors layer-wise weight changes and employs a patience-based mechanism to selectively freeze encoder parameters during local training, significantly reducing client-side computation. CodeCarbon is integrated for end-to-end energy consumption and CO₂-equivalent emission quantification. Contribution/Results: Evaluated across multiple FL architectures, our method reduces training time, total energy consumption, and CO₂eq emissions by up to 23%, while preserving or improving synthesis accuracy—evidenced by minimal MAE variation. This achieves synergistic optimization of sustainability and performance, establishing a novel paradigm for equitable and environmentally sustainable federated medical AI.

Technology Category

• Machine Learning: Efficient ML / Green AI
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Responsible Web: Sustainability and climate impact of web technologies

Abstract

Federated Learning (FL) holds the potential to advance equality in health by enabling diverse institutions to collaboratively train deep learning (DL) models, even with limited data. However, the significant resource requirements of FL often exclude centres with limited computational infrastructure, further widening existing healthcare disparities. To address this issue, we propose a Green AI-oriented adaptive layer-freezing strategy designed to reduce energy consumption and computational load while maintaining model performance. We tested our approach using different federated architectures for Magnetic Resonance Imaging (MRI)-to-Computed Tomography (CT) conversion. The proposed adaptive strategy optimises the federated training by selectively freezing the encoder weights based on the monitored relative difference of the encoder weights from round to round. A patience-based mechanism ensures that freezing only occurs when updates remain consistently minimal. The energy consumption and CO2eq emissions of the federation were tracked using the CodeCarbon library. Compared to equivalent non-frozen counterparts, our approach reduced training time, total energy consumption and CO2eq emissions by up to 23%. At the same time, the MRI-to-CT conversion performance was maintained, with only small variations in the Mean Absolute Error (MAE). Notably, for three out of the five evaluated architectures, no statistically significant differences were observed, while two architectures exhibited statistically significant improvements. Our work aligns with a research paradigm that promotes DL-based frameworks meeting clinical requirements while ensuring climatic, social, and economic sustainability. It lays the groundwork for novel FL evaluation frameworks, advancing privacy, equity and, more broadly, justice in AI-driven healthcare.

Problem

• Reduces energy consumption in federated learning for medical imaging
• Maintains model performance while lowering computational resource needs
• Addresses healthcare disparities by enabling resource-limited institutions to participate

Innovation

• Adaptive encoder freezing reduces FL energy consumption
• Patience-based mechanism freezes weights when updates minimal
• Strategy maintains MRI-to-CT conversion performance while cutting emissions