Syllable-based Korean named entity recognition using convolutional neural network
1
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
Named entity recognition (NER) finds object names in documents or sentences and classifies them into given categories. However, recognizing entity names in natural-language sentences is challenging, as it requires an understanding of various contexts. Recently, many researches have tried to apply deep learning methods to NER and have improved performance. Particularly, the bidirectional long-short-term-memory with conditional random field (bi-LSTM-CRF), which is a recurrent neural network model, is considered the most accurate for NER, as it considers contexts of both directions and is not affected by gradient-vanishing. However, the sequential nature of bi-LSTM-CRF makes the model extremely slow in training and classifying. To overcome this issue of computational speed, we propose a syllable-based Korean NER method using a convolutional neural network with CRF (CNN-CRF). The experiment with three corpora shows that the proposed model achieves a similar level of performance with bi- LSTM-CRF (0.4% improvement); however, it is 27.5% faster than bi-LSTM-CRF.
Abstract
Named entity recognition (NER) finds object names in documents or sentences and classifies them into given categories. However, recognizing entity names in natural-language sentences is challenging, as it requires an understanding of various contexts. Recently, many researches have tried to apply deep learning methods to NER and have improved performance. Particularly, the bidirectional long-short-term-memory with conditional random field (bi-LSTM-CRF), which is a recurrent neural network model, is considered the most accurate for NER, as it considers contexts of both directions and is not affected by gradient-vanishing. However, the sequential nature of bi-LSTM-CRF makes the model extremely slow in training and classifying. To overcome this issue of computational speed, we propose a syllable-based Korean NER method using a convolutional neural network with CRF (CNN-CRF). The experiment with three corpora shows that the proposed model achieves a similar level of performance with bi- LSTM-CRF (0.4% improvement); however, it is 27.5% faster than bi-LSTM-CRF.
참고문헌 (17)
- Y. H. Shin, "Bidirectional LSTM-RNNs-CRF for named entity recognition in Korean", Proceedings of the 28th Annual Conference on Human and Cognitive Language Technology, pp.340~341 (2016)
- D. Y. Lee, "Korean entity recognition system using Bi-directional LSTM-CNN-CRF", Proceedings of the 29th Annual Conference on Human and Cognitive Language Technology, pp.327~329 (2017)
- H. M. Cho, "Named entity recognition using recurrent neural network and convolutional neural network", Proceedings of the KIISE Korea Computer Congress, 2017, pp.636~638 (2017)
- Y. J. Jang, "Named-entity recognition using stacked Bi-LSTM-CRF ensemble model", Proceedings of the KIISE Korea Computer Congress, 2018, pp.2049~2051 (2018)
- S. J. Kwon, "A Korean named entity recognizer using weighted voting based ensemble technique", Proceedings of the 28th Annual Conference on Human and Cognitive Language Technology, pp.333~335 (2016)
- 유홍연, "Bidirectional LSTM CRF 기반의 개체명 인식을 위한 단어 표상의 확장", 정보과학회논문지, 44(3), pp.306~313 (2017), 한국정보과학회
- C. Y. Song, "Named-entity recognizer based on bidirectional LSTM CRFs using improved word embedding models and dictionaries", Proceedings of the KIISE Korea Computer Congress, 2017, pp.699~701 (2017)
- S. H. Na, "Character-based LSTM CRFs for named entity recognition", Proceedings of the KIISE Korea Computer Congress, 2016, pp.729~731 (2016)
- J. W. Min, "Lexicon feature infused characterbased LSTM CRFs for Korean named entity recognition", Proceedings of the 28th Annual Conference on Human and Cognitive Language Technology, pp.99~101 (2016)
- M. A. Cheon, "Character-aware neural networks with multi-head attention mechanism for multilingual named entity recognition", Proceedings of the 30th Annual Conference on Human and Cognitive Language Technology, pp.167~170 (2018)
- 천민아, "Evaluating and applying deep learning-based multilingual named entity recognition", 한국마린엔지니어링학회지, 42(2), pp.106~113 (2018), 한국마린엔지니어링학회[DOI]
- C. K. Lee, "Named entity recognition using deep learning", Proceedings of the 41st KIISE Winter Conference, pp.423~425 (2014)
- C. K. Lee, "Named entity recognition using long short-term memory based recurrent neural network", Proceedings of the KIISE Korea Computer Congress, 2015, pp.645~647 (2015)
- K. H. Choi, "Bio-NER using LSTM-CRF", Proceedings of the 27th Annual Conference on Human and Cognitive Language Technology, pp.85~89 (2015)
- Y. J. Jang, "Named-entity recognition using stacked Bi-LSTM-CRF ensemble model", Proceedings of the KIISE Korea Computer Congress 2018(KCC 2018), pp.2049~2051 (2018)
- "National Institute of Korean Language Named Entity Recognition Corpus"[URL]
- "Korea Maritime and Ocean University Named Entity Recognition Corpus"[URL]