FedAFD: Multimodal Federated Learning via Adversarial Fusion and Distillation
#FedAFD #multimodal federated learning #adversarial fusion #distillation #data privacy #model integration #feature alignment
📌 Key Takeaways
- FedAFD introduces a multimodal federated learning framework using adversarial fusion and distillation.
- The method enhances data privacy by training models locally without sharing raw data.
- It improves model performance by effectively integrating diverse data types across clients.
- Adversarial techniques help align feature distributions from different modalities.
📖 Full Retelling
arXiv:2603.04890v1 Announce Type: cross
Abstract: Multimodal Federated Learning (MFL) enables clients with heterogeneous data modalities to collaboratively train models without sharing raw data, offering a privacy-preserving framework that leverages complementary cross-modal information. However, existing methods often overlook personalized client performance and struggle with modality/task discrepancies, as well as model heterogeneity. To address these challenges, we propose FedAFD, a unified
🏷️ Themes
Federated Learning, Multimodal AI
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
--> Computer Science > Machine Learning arXiv:2603.04890 [Submitted on 5 Mar 2026] Title: FedAFD: Multimodal Federated Learning via Adversarial Fusion and Distillation Authors: Min Tan , Junchao Ma , Yinfu Feng , Jiajun Ding , Wenwen Pan , Tingting Han , Qian Zheng , Zhenzhong Kuang , Zhou Yu View a PDF of the paper titled FedAFD: Multimodal Federated Learning via Adversarial Fusion and Distillation, by Min Tan and 8 other authors View PDF HTML Abstract: Multimodal Federated Learning enables clients with heterogeneous data modalities to collaboratively train models without sharing raw data, offering a privacy-preserving framework that leverages complementary cross-modal information. However, existing methods often overlook personalized client performance and struggle with modality/task discrepancies, as well as model heterogeneity. To address these challenges, we propose FedAFD, a unified MFL framework that enhances client and server learning. On the client side, we introduce a bi-level adversarial alignment strategy to align local and global representations within and across modalities, mitigating modality and task gaps. We further design a granularity-aware fusion module to integrate global knowledge into the personalized features adaptively. On the server side, to handle model heterogeneity, we propose a similarity-guided ensemble distillation mechanism that aggregates client representations on shared public data based on feature similarity and distills the fused knowledge into the global model. Extensive experiments conducted under both IID and non-IID settings demonstrate that FedAFD achieves superior performance and efficiency for both the client and the server. Comments: Accepted by CVPR 2026 Subjects: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2603.04890 [cs.LG] (or arXiv:2603.04890v1 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2603.04890 Focus to learn more arXiv-is...
Read full article at source