Unsupervised cross-domain named entity recognition using entity-aware adversarial training

The success of neural network based methods in named entity recognition (NER) is heavily relied on abundant manual labeled data. However, these NER methods are unavailable when the data is fully-unlabeled in a new domain. To address the problem, we propose an unsupervised cross domain model which leverages labeled data from source domain to predict entities in unlabeled target domain. To relieve the distribution divergence when transferring knowledge from source to target domain, we apply adversarial training. Furthermore, we design an entity-aware attention module to guide the adversarial training to reduce the discrepancy of entity features between different domains. Experimental results demonstrate that our model outperforms other methods and achieves state-of-the-art performance. (c) 2020 Published by Elsevier Ltd.

Automated summary

An unsupervised cross domain model is proposed in this study, which leverages labeled data from the source domain to predict entities in the target domain by applying adversarial training and an entity-aware attention module to reduce feature discrepancy between different domains.