Adversarial training for multi-context joint entity and relation extraction
This work addresses the need for robust extraction models in various domains and languages, but it is incremental as it builds on existing adversarial training methods.
The paper tackled the problem of improving joint entity and relation extraction by applying adversarial training to a baseline model, resulting in state-of-the-art effectiveness across multiple datasets in news, biomedical, and real estate contexts for English and Dutch.
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).