Comparing CNN and LSTM character-level embeddings in BiLSTM-CRF models for chemical and disease named entity recognition

Zenan Zhai, Dat Quoc Nguyen, Karin Verspoor


Abstract
We compare the use of LSTM-based and CNN-based character-level word embeddings in BiLSTM-CRF models to approach chemical and disease named entity recognition (NER) tasks. Empirical results over the BioCreative V CDR corpus show that the use of either type of character-level word embeddings in conjunction with the BiLSTM-CRF models leads to comparable state-of-the-art performance. However, the models using CNN-based character-level word embeddings have a computational performance advantage, increasing training time over word-based models by 25% while the LSTM-based character-level word embeddings more than double the required training time.
Anthology ID:
W18-5605
Volume:
Proceedings of the Ninth International Workshop on Health Text Mining and Information Analysis
Month:
October
Year:
2018
Address:
Brussels, Belgium
Editors:
Alberto Lavelli, Anne-Lyse Minard, Fabio Rinaldi
Venue:
Louhi
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
38–43
Language:
URL:
https://aclanthology.org/W18-5605
DOI:
10.18653/v1/W18-5605
Bibkey:
Cite (ACL):
Zenan Zhai, Dat Quoc Nguyen, and Karin Verspoor. 2018. Comparing CNN and LSTM character-level embeddings in BiLSTM-CRF models for chemical and disease named entity recognition. In Proceedings of the Ninth International Workshop on Health Text Mining and Information Analysis, pages 38–43, Brussels, Belgium. Association for Computational Linguistics.
Cite (Informal):
Comparing CNN and LSTM character-level embeddings in BiLSTM-CRF models for chemical and disease named entity recognition (Zhai et al., Louhi 2018)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-4/W18-5605.pdf