Distantly Supervised Relation Extraction in Federated Settings

TL;DR

This paper introduces a federated learning approach for relation extraction that addresses privacy concerns and label noise, enabling effective training across distributed platforms without centralized data collection.

Contribution

It proposes a novel federated denoising framework with multiple instance learning to improve relation extraction accuracy in distributed, privacy-sensitive environments.

Findings

Effective noise suppression in federated relation extraction

Improved performance on NYT and miRNA datasets

Demonstrates feasibility of federated learning for relation extraction

Abstract

This paper investigates distantly supervised relation extraction in federated settings. Previous studies focus on distant supervision under the assumption of centralized training, which requires collecting texts from different platforms and storing them on one machine. However, centralized training is challenged by two issues, namely, data barriers and privacy protection, which make it almost impossible or cost-prohibitive to centralize data from multiple platforms. Therefore, it is worthy to investigate distant supervision in the federated learning paradigm, which decouples the model training from the need for direct access to the raw data. Overcoming label noise of distant supervision, however, becomes more difficult in federated settings, since the sentences containing the same entity pair may scatter around different platforms. In this paper, we propose a federated denoising framework to suppress label noise in federated settings. The core of this framework is a multiple instance learning based denoising method that is able to select reliable instances via cross-platform collaboration. Various experimental results on New York Times dataset and miRNA gene regulation relation dataset demonstrate the effectiveness of the proposed method.