Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

Enhancing Robustness to Missing Modalities through Clustered Federated Learning

Created by
  • Haebom

Author

Lishan Yang, Wei Emma Zhang, Quan Z. Sheng, Weitong Chen, Lina Yao, Weitong Chen, Ali Shakeri

Outline

In this paper, we propose a novel framework, MMiC, to address the missing modality problem in multimodal federated learning (MFL), a distributed learning method that uses data of various modalities. MMiC mitigates the impact of missing modalities by replacing partial parameters of client models within a cluster, optimizes client selection using the Banzhaf Power Index, and dynamically controls global aggregation using Markovitz Portfolio Optimization. Experimental results show that MMiC outperforms existing federated learning architectures in both global and personalized performance on multimodal datasets with missing modalities.

Takeaways, Limitations

•
Takeaways:
◦
An Effective Solution to the Missing Modality Problem in Multimodal Federated Learning
◦
Improved global and personalized performance through the MMiC framework
◦
Effective application of Banzhaf Power Index and Markovitz Portfolio Optimization to federated learning
◦
Increasing the practicality of federated learning using multimodality data
•
Limitations:
◦
The performance of the proposed framework may depend on specific datasets and experimental settings.
◦
The complexity of MMiC may increase computational costs.
◦
Generalization performance evaluation for various types of modality omission patterns is needed.
◦
Further research is needed on applicability and scalability in real environments.
👍
Made with Slashpage