New ProMoE-FL framework tackles missing data in multimodal federated learning

Researchers have developed ProMoE-FL, a new framework for multimodal federated learning that addresses the challenge of missing data modalities. This approach utilizes a client-aware prototype bank to capture modality priors across different institutions, enabling dynamic synthesis of missing features through a Mixture of Experts model. ProMoE-FL has demonstrated superior performance compared to existing methods on four public chest X-ray datasets, performing well in both homogeneous and heterogeneous settings. AI

IMPACT This research could improve the accuracy and robustness of AI models in healthcare by enabling them to handle incomplete data more effectively.

RANK_REASON The cluster contains a research paper detailing a new framework for multimodal federated learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• CheXpert
• MIMIC-CXR
• NIH Open-I
• ProMoE-FL

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →