Federated Learning with Positive and Unlabeled Data

TL;DR

This paper introduces FedPU, a federated learning framework that effectively learns from positive and unlabeled data across multiple clients with unknown negative classes, outperforming traditional methods.

Contribution

The paper proposes a novel federated learning framework, FedPU, to handle positive and unlabeled data with multiple unknown negative classes, including theoretical analysis and empirical validation.

Findings

FedPU outperforms conventional federated learning methods.

Theoretical analysis provides generalization bounds for FedPU.

Empirical results demonstrate significant performance improvements.

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. Code is available at https://github.com/littleSunlxy/FedPU-torch