计算机科学 ›› 2025, Vol. 52 ›› Issue (5): 227-234.doi: 10.11896/jsjkx.240400035

• 人工智能 • 上一篇    下一篇

融合定长Seq2Seq网络的中文成语智能纠错模型

何春辉, 葛斌, 张翀, 徐浩   

  1. 国防科技大学大数据与决策实验室 长沙 410073
  • 收稿日期:2024-04-04 修回日期:2024-08-20 出版日期:2025-05-15 发布日期:2025-05-12
  • 通讯作者: 葛斌(gebin@nudt.edu.cn)
  • 作者简介:(xtuhch@163.com)
  • 基金资助:
    国家重点研发计划(2022YFB3103600)

Intelligent Error Correction Model for Chinese Idioms Fused with Fixed-length Seq2Seq Network

HE Chunhui, GE Bin, ZHANG Chong, XU Hao   

  1. Laboratory for Big Data and Decision,National University of Defense Technology,Changsha 410073,China
  • Received:2024-04-04 Revised:2024-08-20 Online:2025-05-15 Published:2025-05-12
  • About author:HE Chunhui,born in 1991,master.His main research interests include information processing and artificial intelligence.
    GE Bin,born in 1979,Ph.D,researcher.His main research interests include big data analysis and information extraction.
  • Supported by:
    National Key Research and Development Program of China(2022YFB3103600).

摘要: 四字成语作为一类特殊词语,在中文使用中非常流行。随着中文纠错任务的发展,中文成语的智能纠错已经成为自然语言处理领域的一个研究热点。针对现有方法在中文成语智能纠错任务上准确率偏低的问题,提出了一种融合定长Seq2Seq网络的中文成语智能纠错模型。它在底层通过融合Seq2Seq网络架构和注意力机制,并结合混合数据集构造方法,共同训练得到输入和输出端序列长度固定的Seq2Seq模型,用来完成中文四字成语智能纠错任务。在大型公开中文成语纠错数据集上的实验结果表明,定长Seq2Seq模型优于现有方法,能够实现同一个模型同时兼容乱序、缺字和错字3种不同的中文成语智能纠错目标。它的综合纠错准确率可以达到91.3%,比最优基线模型高出11.73%。

关键词: 成语纠错, 定长Seq2Seq, 双向GRU, 注意力机制

Abstract: As a special kind of words,four-character idioms are very popular in Chinese.With the development of Chinese error correction task,intelligent error correction for Chinese idioms has become a research hotspot in natural language processing(NLP) domain.For the low accuracy of the existing methods in intelligent error correction task for Chinese idioms,this paper proposes an intelligent error correction model for Chinese idioms fused with fixed-length Seq2Seq network.In the bottom layer,Seq2Seq network architecture and attention mechanism are combined with hybrid dataset construction method to train Seq2Seq model with fixed input and output sequence length,which is used to solve intelligent error correction task for Chinese four-character idioms.Experimental results on a large public Chinese idiom error correction dataset show that the performance of fixed-length Seq2Seq model is better than the existing methods,and it can achieve the goal of intelligent error correction of three diffe-rent Chinese idioms:out-of-order,missing character and wrong character.Its comprehensive error correction accuracy can reach 91.3%,which is 11.73% higher than the optimal baseline model.

Key words: Idioms error correction, Fixed length Seq2Seq, BiGRU, Attention mechanism

中图分类号: 

  • TP391
[1]XU H,HE C H,ZHANG C,et al.A Multi-channel ChineseText Correction Method Based on Grammatical Error Diagnosis[C]//2022 8th International Conference on Big Data and Information Analytics(BigDIA).2022:396-401.
[2]SUN Q J,LIANG J G,LI S,Chinese grammatical error correction model based on bidirectional and auto-regressive transfor-mers noiser[J].Journal of Computer Applications,2022,42(3):860-866.
[3]YOO Y.An Analysis on Four-character idiom in the Contempo-rary Chinese Dictionary[J].Journal of Chinese Humanities,2010(46):93-109.
[4]WANG Y,WANG Y,DANG K,et al.A comprehensive survey of grammatical error correction[J].ACM Transactions on Intelligent Systems and Technology(TIST),2021,12(5):1-51.
[5]WU C H,LIU C H,HARRIS M,et al.Sentence correction incorporating relative position and parse template language models[J].IEEE Transactions on Audio Speech & Language Proces-sing,2010,18(6):1170-1181.
[6]YU C H,CHEN H H.Detecting word ordering errors in Chi-nese sentences for learning Chinese as a foreign language[C]//Proceedings of COLING.2012.
[7]CHENG S M,YU C H,CHEN H H.Chinese Word Ordering Errors Detection and Correction for Non-Native Chinese Language Learners[C]//The 25th International Conference on Computational Linguistics.2014:279-289.
[8]FU K,HUANG J,DUAN Y.Youdao's winning solution to the NLPCC-2018 task 2 challenge:a neural machine translation approach to Chinese grammatical error correction[C]//NLPCC2018.Cham:Springer,2018:341-350.
[9]SALHAB M,ABU-KHZAM F.AraSpell:A Deep Learning Approach for Arabic Spelling Correction[J].arXiv:2405.06981,2024.
[10]HUANG Y,ZENG Q,LEI Q,et al.Smartphone heading correc-tion method based on LSTM neural network[C]//China Satellite Navigation Conference.Singapore:Springer Nature Singapore,2022:415-425.
[11]ZHANG C,JIANG D,GAO Y,et al.A hierarchical tensor error correction GRU model[J].Information Sciences,2023,642:119156.
[12]WANG N,LI Z.Short term power load forecasting based onBES-VMD and CNN-Bi-LSTM method with error correction[J].Frontiers in Energy Research,2023,10:1076529.
[13]LI J,GUO J,ZHU Y,et al.Sequence-to-action:Grammatical error correction with action guided sequence generation[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:10974-10982.
[14]ZHU C,YING Z,ZHANG B,et al.MDCSpell:A multi-task detector-corrector framework for Chinese spelling correction[C]//Findings of the Association for Computational Linguistics:ACL 2022.2022:1244-1253.
[15]HOKAMP C,LIU Q.Lexically Constrained Decoding for Se-quence Generation Using Grid Beam Search[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics(ACL).Vancouver,Canada,2017:1535-1546.
[16]SUTSKEVER I,VINYALS O,LE Q V.Sequence to sequence learning with neural networks[J].arXiv:1409.3215,2014.
[17]CHO K,MERRIENBOER B,GULCEHRE C,et al.LearningPhrase Representations using RNN Encoder-Decoder for Statistical Machine Translation[C]//EMNLP.2014.
[18]GEMECHU E,KANAGACHIDAMBARESAN G R.Text-Text Neural Machine Translation:A Survey[J].Optical Memory and Neural Networks,2023,32(2):59-72.
[19]DAS B,MAJUMDER M,PHADIKAR S,et al.Automatic question generation and answer assessment:a survey[J/OL].https://telrp.springeropen.com/counter/pdf/10.1186/s41039-021-00151-1.pdf.
[20]ZHAO S,LI Q,HE T J,et al.A Step-by-Step Gradient Penalty with Similarity Calculation for Text Summary Generation[J].Neural Processing Letters,2022,55(4):4111-4126.
[21]BAHDANAU D,CHO K,BENGIO Y.Neural machine translation by jointly learning to align and translate[J].arXiv:1409.0473,2014.
[22]LUONG M T,PHAM H,MANNING C D.Effective approaches to attention-based neural machine translation[J].arXiv:1508.04025,2015.
[23]PODDA M,BONECHI S,PALLADINO A,et al.Classification of Neisseria meningitidis genomes with a bag-of-words approach and machine learning[J].iScience,2024,27(3):1-15.
[24]WANG H,KUROSAWA M,KATSUMATA S,et al.Chinese grammatical correction using BERT-based pre-trained model[J].arXiv:2011.02093,2020.
[25]XU M.Pycorrector:Text error correction tool [EB/OL].(2024-02-03).https://github.com/shibing624/pycorrector.
[26]CUI Y,CHE W,LIU T,et al.Pre-training with whole word masking for Chinese bert[J].IEEE/ACM Transactions on Audio,Speech,and Language Processing,2021,29:3504-3514.
[27]CZHANG Y,LI Z,BAO Z,et al.MuCGEC:a Multi-ReferenceMulti-Source Evaluation Dataset for Chinese Grammatical Error Correction[J].arXiv:2204.10994,2022.
[28]TIAN T,SONG C,TING J,et al.A French-to-English Machine Translation Model Using Transformer Network[J].Procedia Computer Science,2022,199:1438-1443.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!