Paper Accepted: CAR-MFL: Cross-Modal Augmentation by Retrieval for Multimodal Federated Learning with Missing Modalities

The paper “CAR-MFL: Cross-Modal Augmentation by Retrieval for Multimodal Federated Learning with Missing Modalities” by Pranav Poudel, Prashant Shrestha, Sanskar Amgain, Yash Raj Shrestha, Prashnna Gyawali, Binod Bhattarai has be accepted  at MICCAI 2024.

This paper presents a retrieval-based cross-modal data augmentation approach for multi-modal federated learning with missing modalities by leveraging publicly available multi-modal datasets. The simple yet effective method surpasses previous federated learning frameworks while requiring significantly fewer public data samples.

🔗Preprint: https://lnkd.in/dezmNUFn