学术文献的摘要是对文献主要内容的浓缩,摘要不同部分的语步具有不同的信息,语步的自动识别和抽取对于学术摘要的后续研究有着重要的应用价值,而目前语步识别的研究相对较少,并且相关算法的效果还需要提高。针对上述问题,该文提出了一种基于ERNIE-BiGRU模型的语步识别算法。该算法首先结合中文句法分析理论提出基于句法依存关系的多语步结构拆分法,对学术文献摘要多语步结构进行自动拆分,获得多个单语步结构;然后构建用于训练的单语步结构语料库,并利用知识增强语义表示预训练模型,训练出句子级词向量;最后将训练出的单语步结构词向量信息输入双向门限循环单元(BiGRU)进行摘要语步自动化识别,取得了良好的效果。实验结果表明,该算法具有较好的鲁棒性和较高的识别精度,在结构化和非结构化摘要上的识别准确率分别达到了96.57%和93.75%。
Abstract
The academic abstract summarizes key points in a research paper, with a series of moves conveying different information. The automatic recognition and extraction of moves could provide a valuable foundation for other tasks related with the academic abstract. This paper proposed a move recognition algorithm for academic abstract based on ERNIE-BiGRU model. Firstly, a multi-move structure splitting method based on dependency structure is proposed, identifying the multiple single-move structure in the academic abstract. Secondly, a single-move structure corpus is constructed, and a pre-training model of knowledge-enhanced semantic representation is employed to train sentence-level word vectors. Finally, the trained word vector with single move structure information is input into BiGRU for automatic recognition of moves. The experimental results show that the algorithm has good robustness and high recognition accuracy, achieving 96.57% and 93.75% recognition accuracy for structured and unstructured abstracts, respectively.
关键词
中文句法分析 /
多语步结构拆分 /
ERNIE-BiGRU模型
{{custom_keyword}} /
Key words
Chinese parsing /
multi-move structure splitting /
ERNIE-BiGRU model
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 沈思,胡昊天,叶文豪,等.基于全字语义的摘要结构功能自动识别研究[J].情报学报,2019.38(1): 79-88.
[2] 王末,崔运鹏,陈丽,等.基于深度学习的学术论文语步结构分类方法研究[J].数据分析与知识发现,2020.4(6): 60-68.
[3] LISTED N.A proposal for more informative abstracts of clinical articales. Ad Hoc working group for critical appraisal of the medical literature[J].Annals of Internal Medicine,1987,106(4): 598-604.
[4] HARTLEY J. Current findings from research on struct abstract[J].Med Libr Assoc,2014,92(3): 368-371.
[5] 白光祖,何远标,马建霞,等.利用小样本量机器学习实现学术文摘结构的自动识别[J].现代图书情报,2014.30(7): 34-40.
[6] ZHANG Z X,LIU H,DING L P,et al.Moves recognition in abstract of research paper based on deep learning[C]//Proceeding of ACM/IEEE Joint Conference on Digital Libraries, Champaign(US),2019: 390-391.
[7] 岳增营,叶霞,刘睿珩.基于语言模型的预训练技术研究综述[J] .中文信息学报,2021,35(09): 15-29.
[8] CHENG X,ZHANG C,LI Q X.Improved Chinese short text classification based on ERNIE—BiGRU Model[J].Journal of Physics: Conference Series, 2021(6): 1-11.
[9] 叶娜,黎天宇,蔡东风,等.利用依存句法关系改进神经译文质量估计[J].中文信息学报,2021,35(09): 46-57.
[10] SUN Y,WANG S H,LI Y K, et al. ERNIE: Enhanced representation through knowledge integration[EB/OL].https://arxiv.org/abs/1904.09223[2019-12-23].
[11] YU G,ZHANG Z,LIU H,et al.Masked sentence model based on BERT for move recognition in medical scientific abstracts[J]. Journal of Data and Information Science, 2019(04): 42-55.
[12] JACOB D, CHANG M W, KENTON L, et al. BERT: Pre-training of deep bidirectional transformers for language understanding[C]//Proceeding of Conference on the North American Chapter of the Association for Computational Linguisitcs: Human Language Technologies. Minneapolis, Minnesota: Association for Computational Linguistics,2019: 4171-4186.
[13] DAI J, CHEN C. Text classification system of academic papers based on hybrid Bert-BiGRU model[C]//Proceeding of International Conference on Intelligent Human-Machine Systems and Cybernetics. Hangzhou, China: IEEE,2020: 40-44.
[14] 屠可伟,李俊.句法分析前沿动态综述[J].中文信息学报,2020,34(07): 30-41.
[15] 黄彤,李斌,闫培艺,等.基于抽象语义表示的汉语构式标注与分析[J].中文信息学报,2020,34(10): 1-9.
[16] 贾延延,程学旗,冯键.基于LSTM的层次化篇章依存分析方法[J].中文信息学报,2021,35(01): 1-8.
[17] 许晨彬.基于BERT的深层语义特征提取在复句层次结构分析中的应用[D].武汉: 华中师范大学硕士学位论文,2020.
[18] 郑梦悦,秦春秀,马续补,等.面向中文科技文献非结构化摘要的知识元表示与抽取研究[J].情报理论与实践,2020,43(2): 157-163.
[19] WANG R P,ZHANG C Z,ZHANG Y Y,et al.Extracting methodological sentences from unstructured abstracts of academic articles[C]//Proceeding of International Conference on Sustainable Digital Communities.2020: 790-798.
[20] 周志超.中文图情期刊摘要的核心要素与逻辑结构分析[J].情报科学,2018,36(3): 8-12.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
国家自然科学基金(71673213)
{{custom_fund}}
PDF(5503 KB)
