Adversarial training for multi-context joint entity and relation extraction

TL;DR

This paper demonstrates that adversarial training enhances the robustness and effectiveness of joint entity and relation extraction models across multiple datasets, languages, and contexts.

Contribution

It introduces the application of adversarial training to a general baseline model for joint entity and relation extraction, achieving state-of-the-art results.

Findings

Improved accuracy on news, biomedical, and real estate datasets.

Effective across English and Dutch languages.

Enhanced robustness of extraction models.

Abstract

Adversarial training (AT) is a regularization method that can be used to improve the robustness of neural network methods by adding small perturbations in the training data. We show how to use AT for the tasks of entity recognition and relation extraction. In particular, we demonstrate that applying AT to a general purpose baseline model for jointly extracting entities and relations, allows improving the state-of-the-art effectiveness on several datasets in different contexts (i.e., news, biomedical, and real estate data) and for different languages (English and Dutch).