Communication-Efficient and Robust Multi-Modal Federated Learning via Latent-Space Consensus
Mohamed Badi, Chaouki Ben Issaid, Mehdi Bennis
This paper presents CoMFed, a new method for federated learning (training AI models across multiple devices without sharing raw data) that handles situations where different devices have different types of data and models. The key innovation is using compressed representations and alignment techniques to help devices work together efficiently while keeping communication costs low and protecting privacy.
federated learningmulti-modal learningcommunication efficiencyprivacy-preserving ML