HybridVFL: Advancing Federated Learning for Multimodal Data at the Edge

This research explores a novel approach to vertical federated learning, crucial for privacy-preserving multimodal classification in edge computing environments. The disentangled feature learning strategy likely enhances performance while addressing challenges related to data heterogeneity and communication overhead.

• •Addresses the challenge of privacy-preserving multimodal classification.
• •Focuses on a federated learning approach, beneficial for distributed data.
• •Explores feature disentanglement for improved performance and efficiency.