IEEE transactions on image processing : a publication of the IEEE Signal Processing Society
Token-Level Prompt Mixture with Parameter-Free Routing for Federated Domain Generalization.
Shuai Gong, Chaoran Cui, Xiaolin Dong, Xiushan Nie, Lei Zhu, Xiaojun Chang
Published: 202610.1109/TIP.2026.3652431
Abstract
Domain Generalization (FedDG) aims to train a globally generalizable model on data from decentralized, heterogeneous clients. While recent work has adapted vision-language models for FedDG using prompt learning, the prevailing "one-prompt-fits-all" p…
Preview only. Read the full abstract at the source