Generative Autoregressive Transformers for Model-Agnostic Federated MRI Reconstruction

2025-02-06

AI Summary

Existing federated MRI reconstruction methods assume homogeneous model architectures across participating sites, limiting adaptability to heterogeneous computational resources and clinical requirements. To address this, we propose the first privacy-preserving federated learning framework supporting architectural heterogeneity. Our approach comprises three key components: (1) a model-agnostic generative federated paradigm; (2) a site-specific prompt-guided autoregressive Transformer (GAT) prior enabling controllable, cross-center image synthesis; and (3) a synergistic training mechanism integrating federated knowledge distillation with heterogeneous model coordination. Evaluated on a multi-center MRI dataset, our method significantly outperforms state-of-the-art federated baselines—maintaining high local reconstruction fidelity while substantially improving cross-site generalization. This work establishes a new paradigm for flexible, secure, and clinically adaptable federated medical imaging collaboration.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Although learning-based models hold great promise for MRI reconstruction, single-site models built on limited local datasets often suffer from poor generalization. This challenge has spurred interest in collaborative model training on multi-site datasets via federated learning (FL) -- a privacy-preserving framework that aggregates model updates instead of sharing imaging data. Conventional FL builds a global model by aggregating locally trained model weights, inherently constraining all sites to a homogeneous model architecture. This rigid homogeneity requirement forces sites to forgo architectures tailored to their compute infrastructure and application-specific demands. Consequently, existing FL methods for MRI reconstruction fail to support model-heterogeneous settings, where individual sites are allowed to use distinct architectures. To overcome this fundamental limitation, here we introduce FedGAT, a novel model-agnostic FL technique based on generative autoregressive transformers. FedGAT decentralizes the training of a global generative prior that captures the distribution of multi-site MR images. For enhanced fidelity, we propose a novel site-prompted GAT prior that controllably synthesizes MR images from desired sites via autoregressive prediction across spatial scales. Each site then trains its site-specific reconstruction model -- using its preferred architecture -- on a hybrid dataset comprising the local MRI dataset and GAT-generated synthetic MRI datasets for other sites. Comprehensive experiments on multi-institutional datasets demonstrate that FedGAT supports flexible collaborations while enjoying superior within-site and across-site reconstruction performance compared to state-of-the-art FL baselines.

Problem

• MRI reconstruction generalization
• Model-heterogeneous federated learning
• Generative autoregressive transformers

Innovation

• Generative autoregressive transformers for MRI
• Model-agnostic federated learning technique
• Site-prompted GAT prior synthesis