Federated Learning with Positive and Unlabeled Data

Xinyang Lin, Hanting Chen, Yixing Xu, Chao Xu, Xiaolin Gui, Yiping Deng, Yunhe Wang
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:13344-13355, 2022.

Abstract

We study the problem of learning from positive and unlabeled (PU) data in the federated setting, where each client only labels a little part of their dataset due to the limitation of resources and time. Different from the settings in traditional PU learning where the negative class consists of a single class, the negative samples which cannot be identified by a client in the federated setting may come from multiple classes which are unknown to the client. Therefore, existing PU learning methods can be hardly applied in this situation. To address this problem, we propose a novel framework, namely Federated learning with Positive and Unlabeled data (FedPU), to minimize the expected risk of multiple negative classes by leveraging the labeled data in other clients. We theoretically analyze the generalization bound of the proposed FedPU. Empirical experiments show that the FedPU can achieve much better performance than conventional supervised and semi-supervised federated learning methods.

Cite this Paper


BibTeX
@InProceedings{pmlr-v162-lin22b, title = {Federated Learning with Positive and Unlabeled Data}, author = {Lin, Xinyang and Chen, Hanting and Xu, Yixing and Xu, Chao and Gui, Xiaolin and Deng, Yiping and Wang, Yunhe}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {13344--13355}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/lin22b/lin22b.pdf}, url = {https://proceedings.mlr.press/v162/lin22b.html}, abstract = {We study the problem of learning from positive and unlabeled (PU) data in the federated setting, where each client only labels a little part of their dataset due to the limitation of resources and time. Different from the settings in traditional PU learning where the negative class consists of a single class, the negative samples which cannot be identified by a client in the federated setting may come from multiple classes which are unknown to the client. Therefore, existing PU learning methods can be hardly applied in this situation. To address this problem, we propose a novel framework, namely Federated learning with Positive and Unlabeled data (FedPU), to minimize the expected risk of multiple negative classes by leveraging the labeled data in other clients. We theoretically analyze the generalization bound of the proposed FedPU. Empirical experiments show that the FedPU can achieve much better performance than conventional supervised and semi-supervised federated learning methods.} }
Endnote
%0 Conference Paper %T Federated Learning with Positive and Unlabeled Data %A Xinyang Lin %A Hanting Chen %A Yixing Xu %A Chao Xu %A Xiaolin Gui %A Yiping Deng %A Yunhe Wang %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-lin22b %I PMLR %P 13344--13355 %U https://proceedings.mlr.press/v162/lin22b.html %V 162 %X We study the problem of learning from positive and unlabeled (PU) data in the federated setting, where each client only labels a little part of their dataset due to the limitation of resources and time. Different from the settings in traditional PU learning where the negative class consists of a single class, the negative samples which cannot be identified by a client in the federated setting may come from multiple classes which are unknown to the client. Therefore, existing PU learning methods can be hardly applied in this situation. To address this problem, we propose a novel framework, namely Federated learning with Positive and Unlabeled data (FedPU), to minimize the expected risk of multiple negative classes by leveraging the labeled data in other clients. We theoretically analyze the generalization bound of the proposed FedPU. Empirical experiments show that the FedPU can achieve much better performance than conventional supervised and semi-supervised federated learning methods.
APA
Lin, X., Chen, H., Xu, Y., Xu, C., Gui, X., Deng, Y. & Wang, Y.. (2022). Federated Learning with Positive and Unlabeled Data. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:13344-13355 Available from https://proceedings.mlr.press/v162/lin22b.html.

Related Material