Chinese Word Segmentation and Recognition Based on Separable Convolution Bidirectional Long Short-Term Memory and Feature Point

Fan Sun; Zijiao Chen; Jingrui  Pei

doi:10.6180/jase.202104_24(2).0014

Chinese Word Segmentation and Recognition Based on Separable Convolution Bidirectional Long Short-Term Memory and Feature Point

Computer Science and Information Engineering

Design and Numerical Simulation of a Novel High-Speed Multi-Degrees of Freedom Piezoelectric positioner

Fan Sun¹ , Zijiao Chen¹ , and Jingrui Pei This email address is being protected from spambots. You need JavaScript enabled to view it.^2,3

¹School of Chinese Studies, Dalian University of Foreign Languages Dalian,116044, China
²Chinese College of Minority Languages and Literature, Minzu University of China Beijing, 100000, China
³Software College, Shenyang Normal University Shenyang, 110034, China

Received: August 1, 2020
Accepted: November 16, 2020
Publication Date: April 1, 2020

Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Download Citation: ||https://doi.org/10.6180/jase.202104_24(2).0014

ABSTRACT

Chinese automatic word segmentation is the premise of Chinese information processing, which is widely used in Chinese full-text retrieval, Chinese automatic full-text translation, Chinese text-to-speech conversion (TTS) and other fields. A dictionary plays an important role in Chinese word segmentation. The advantages and disadvantages of the word segmentation mechanism directly affect the speed and efficiency of Chinese word segmentation. Therefore, we propose a deep learning method for Chinese word segmentation. First, a separable convolution bidirectional long and short-term memory condition random field word segmentation model with feature points containing dictionary features is constructed. The model parameters are obtained by training on the existing word segmentation corpus. Then, the software engineering field text is used as the small-scale word segmentation training corpus. The word segmentation model of general corpus is fine-tuned. The experimental results show that the transfer learning reduces the iteration times of the domain segmentation model. Meanwhile, compared with other Chinese word segmentation models, the proposed model can reduce the corpus labeling time in training process and realize the cross-domain transfer learning of word segmentation model.

Keywords: Chinese word segmentation; deep learning; dictionary feature; bidirectional long and short-term memory; condition random field; separable convolution; Feature point

REFERENCES

[1] Yan Li, Yinghua Zhang, Xiaoping Huang, Xucheng Yin, and Hongwei Hao. Chinese word segmentation with local and global context representation learning. High Technology Letters, 21(1):71–77, 2015.
[2] Ke Zhu. Analysis of Chinese word segmentation technology. In Applied Mechanics and Materials, volume 687-691, pages 1540–1543. Trans Tech Publications Ltd, 2014.
[3] Hong-gang Zhang and Huan Li. Chinese Word Segmentation Method on the Basis of Bidirectional Long-Short Term Memory Model. Journal of South China University of Technology, 45(3):61–67, 2017.
[4] Yue Zhao, Hang Li, Shoulin Yin, and Yang Sun. A new Chinese word segmentation method based on maximum matching. Journal of Information Hiding and Multimedia Signal Processing, 9(6):1528–1535, 2018.
[5] Lichi Yuan. Improvement for the automatic part-ofspeech tagging based on Hidden Markov Model. In ICSPS 2010 - Proceedings of the 2010 2nd International Conference on Signal Processing Systems, volume 1, 2010.
[6] Jinli Che, Liwei Tang, Shijie Deng, and Xujun Su. Chinese word segmentation based on bidirectional GRUCRF model. International Journal of Performability Engineering, 14(12):3066–3075, dec 2018.
[7] Yusi Cheng and Yuntao Shi. Domain Adaption of Chinese Word Segmentation Based on Deep Learning and Transfer Learning. Chinese Journal of Information Processing, 33(9), 2019.
[8] V. P. Maslov and T. V. Maslova. On the boundedness law for the number of words in an overabundant dictionary. Mathematical Notes, 85(1-2):296–301, feb 2009.
[9] Yushi Yao and Zheng Huang. Bi-directional LSTM recurrent neural network for chinese word segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 9950 LNCS, pages 345–353. Springer Verlag, 2016.
[10] Zhang Jian, Cheng RenHong, Wang Kai, and Zhao Hong. Research on born-digital image text extraction based on conditional random field. International Journal of High Performance Systems Architecture, 5(1):39–49, 2014.
[11] Alida H.P.M. de Rooij, Katrien G. Luijkx, Anja G. Declercq, Peggy M.J. Emmerink, and Jos M.G.A. Schols. Professional Caregivers’ Mental Health Problems and Burnout in Small-Scale and Traditional Long Term Care Settings for Elderly People With Dementia in The Netherlands and Belgium. Journal of the American Medical Directors Association, 13(5):486.e7–486.e11, 2012.
[12] J. M. Chen, J. Liu, and E. Huang. Abbreviation Expansion Interpretation Recognition Based on Semisupervised CRF. 39(4):203–209, 2013.
[13] Shoulin Yin, Ye Zhang, and Shahid Karim. Region search based on hybrid convolutional neural network in optical remote sensing images. International Journal of Distributed Sensor Networks, 15(5), may 2019.
[14] Peng Li, Zhikui Chen, Laurence Tianruo Yang, Qingchen Zhang, and M Jamal Deen. Deep convolutional computation model for feature learning on big data in internet of things. IEEE Transactions on Industrial Informatics, 14(2):790–798, 2017.
[15] Ying Huang, Mingqing Hu, Xianguo Yu, Tao Wang, and Chen Yang. Transfer learning of deep neural network for speech emotion recognition. In Communications in Computer and Information Science, volume 663, pages 721–729. Springer Verlag, 2016.
[16] Ravi K Samala, Heang Ping Chan, Lubomir M. Hadjiiski, Mark A Helvie, Kenny H Cha, and Caleb D Richter. Multi-task transfer learning deep convolutional neural network: Application to computer-aided diagnosis of breast cancer on mammograms. Physics in Medicine and Biology, 62(23):8894–8908, 2017.
[17] Rekia Kadari, Yu Zhang, Weinan Zhang, and Ting Liu. CCG supertagging via Bidirectional LSTM-CRF neural architecture. Neurocomputing, 283:31–37, mar 2018.
[18] Junxing Shi, Haiguang Wen, Yizhen Zhang, Kuan Han, and Zhongming Liu. Deep recurrent neural network reveals a hierarchy of process memory during dynamic natural vision. Human Brain Mapping, 39(5):2269–2282, may 2018.
[19] Yubin Dai, Christopher S.G. Khoo, and Teck Ee Loh. A new statistical formula for Chinese text segmentation incorporating contextual information. In Proceedings of the 22nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 1999, pages 82–89. Association for Computing Machinery, Inc, aug 1999.
[20] G E Hinton, N Srivastava, A Krizhevsky, I Sutskever, and R R Salakhutdinov. Improving neural networks by preventing co-adaptation of feature detectors. Technical report.
[21] Xiangdong Li and Cheng Zhang. Research on enhancing the effectiveness of the Chinese text automatic categorization based on ICTCLAS segmentation method. In Proceedings of the IEEE International Conference on Software Engineering and Service Sciences, ICSESS, pages 267–270, 2013.