Leveraging Foundation Models for Multi-modal Federated Learning with Incomplete Modality
This addresses a realistic challenge in federated learning for distributed multi-modal data, though it is incremental as it builds on existing FL and pre-trained model techniques.
The paper tackles the problem of missing modalities in multi-modal federated learning by proposing FedMVP, a method that uses pre-trained models for modality completion and representation transfer, achieving superior performance on image-text classification datasets and robustness to missing data.
Federated learning (FL) has obtained tremendous progress in providing collaborative training solutions for distributed data silos with privacy guarantees. However, few existing works explore a more realistic scenario where the clients hold multiple data modalities. In this paper, we aim to solve a novel challenge in multi-modal federated learning (MFL) -- modality missing -- the clients may lose part of the modalities in their local data sets. To tackle the problems, we propose a novel multi-modal federated learning method, Federated Multi-modal contrastiVe training with Pre-trained completion (FedMVP), which integrates the large-scale pre-trained models to enhance the federated training. In the proposed FedMVP framework, each client deploys a large-scale pre-trained model with frozen parameters for modality completion and representation knowledge transfer, enabling efficient and robust local training. On the server side, we utilize generated data to uniformly measure the representation similarity among the uploaded client models and construct a graph perspective to aggregate them according to their importance in the system. We demonstrate that the model achieves superior performance over two real-world image-text classification datasets and is robust to the performance degradation caused by missing modality.