Class Disjoint Data
Class disjoint data, where different datasets or subsets contain non-overlapping sets of classes, presents significant challenges for machine learning, particularly in federated learning and hyperspectral image classification. Current research focuses on developing algorithms and model architectures, such as manifold reshaping and transformer-based fusion, to address issues like data collapse and negative transfer arising from this data structure. Overcoming these challenges is crucial for improving the accuracy and robustness of machine learning models in diverse applications, ranging from personalized medicine to remote sensing.
Papers
Federated Learning under Partially Class-Disjoint Data via Manifold Reshaping
Ziqing Fan, Jiangchao Yao, Ruipeng Zhang, Lingjuan Lyu, Ya Zhang, Yanfeng Wang
Federated Learning with Bilateral Curation for Partially Class-Disjoint Data
Ziqing Fan, Ruipeng Zhang, Jiangchao Yao, Bo Han, Ya Zhang, Yanfeng Wang