[1]夏林中,叶剑锋,罗德安,等.基于BERT-BiLSTM模型的短文本自动评分系统[J].深圳大学学报理工版,2022,39(3):349-354.[doi:10.3724/SP.J.1249.2022.03349]
 XIA Linzhong,YE Jianfeng,LUO Dean,et al.Short text automatic scoring system based on BERT-BiLSTM model[J].Journal of Shenzhen University Science and Engineering,2022,39(3):349-354.[doi:10.3724/SP.J.1249.2022.03349]
点击复制

基于BERT-BiLSTM模型的短文本自动评分系统
分享到:

《深圳大学学报理工版》[ISSN:1000-2618/CN:44-1401/N]

卷:
第39卷
期数:
2022年第3期
页码:
349-354
栏目:
电子与信息科学
出版日期:
2022-05-16

文章信息/Info

Title:
Short text automatic scoring system based on BERT-BiLSTM model
文章编号:
202203014
作者:
夏林中叶剑锋罗德安管明祥刘俊曹雪梅
深圳信息职业技术学院人工智能技术应用工程实验室,广东深圳 518172
Author(s):
XIA Linzhong YE Jianfeng LUO De’an GUAN Mingxiang LIU Jun and CAO Xuemei
Engineering Applications of Artificial Intelligence Technology Laboratory, Shenzhen Institute of Information Technology, Shenzhen 518172, Guangdong Province, P. R. China
关键词:
信号与信息处理自然语言处理BERT语言模型短文本自动评分长短时记忆网络二次加权kappa系数
Keywords:
signal and information processing natural language processing BERT language model short text automatic scoring long short-term memory net quadratic weighted kappa coefficient
分类号:
TP18;H08
DOI:
10.3724/SP.J.1249.2022.03349
文献标志码:
A
摘要:
针对短文本自动评分中存在的特征稀疏、一词多义及上下文关联信息少等问题,提出一种基于BERT-BiLSTM的短文本自动评分模型.使用BERT语言模型预训练大规模语料库习得通用语言的语义特征,通过预训练好的BERT语言模型预微调下游具体任务的短文本数据集习得短文本的语义特征和关键词特定含义,再通过BiLSTM捕获深层次上下文关联信息,最后将获得的特征向量输入Softmax回归模型进行自动评分.实验结果表明,对比CharCNN、CNN、LSTM和BERT等基准模型,基于BERT-BiLSTM的短文本自动评分模型所获的二次加权kappa系数平均值最好.
Abstract:
A short text automatic scoring model based on BERT-BiLSTM is proposed to aim at solving the problems of sparse features, polysemy and less context related information in short text automatic scoring. Firstly, the large-scale corpus is pre-trained with BERT to acquire the semantic features of the general language. Then the semantic features of short text and the semantics of keywords in a specific context are acquired through the short text data set pre-fined by BERT. And then the deep-seated context dependency is captured through BiLSTM. Finally, the obtained feature vectors are input into Softmax regression model for automatic scoring. The experimental results show that compared with other benchmark models (CharCNN, CNN, LSTM and BERT), the short text automatic scoring model based on BERT-BiLSTM obtains the best average value of quadratic weighted kappa coefficient.

相似文献/References:

[1]刘进忙,倪鹏,李超,等.复杂战场环境下的分坐标处理[J].深圳大学学报理工版,2014,31(2):138.[doi:10.3724/SP.J.1249.2014.02138]
 Liu Jinmang,Ni Peng,Li Chao,et al.The independent coordinate processing in complex battlefield[J].Journal of Shenzhen University Science and Engineering,2014,31(3):138.[doi:10.3724/SP.J.1249.2014.02138]
[2]贺成龙,秦洪,于永生.一种空中目标航迹的自适应跟踪算法[J].深圳大学学报理工版,2014,31(4):361.[doi:10.3724/SP.J.1249.2014.04361]
 He Chenglong,Qin Hong,and Yu Yongsheng.An adaptive tracking algorithm for aerial target[J].Journal of Shenzhen University Science and Engineering,2014,31(3):361.[doi:10.3724/SP.J.1249.2014.04361]
[3]陈星宇,黄俊文,周展,等.基于本体论的大数据下用户需求表征[J].深圳大学学报理工版,2017,34(2):173.
 Chen Xingyu,Huang Junwen,Zhou Zhan,et al. Ontology-based user needs representation in the big data context[J].Journal of Shenzhen University Science and Engineering,2017,34(3):173.
[4]陈星宇,周展,黄俊文,等.基于关键词挖掘的客户细分方法[J].深圳大学学报理工版,2017,34(3):300.[doi:10.3724/SP.J.1249.2017.03300]
 Chen Xingyu,Zhou Zhan,Huang Junwen,et al.A keyword-based mining method for customer segmentation[J].Journal of Shenzhen University Science and Engineering,2017,34(3):300.[doi:10.3724/SP.J.1249.2017.03300]
[5]胡布焕,张晶,张凌.一种基于语义相似的中文文档抄袭检测方法[J].深圳大学学报理工版,2020,37(增刊1):107.[doi:10.3724/SP.J.1249.2020.99107]
 HU Buhuan,ZHANG Jing,and ZHANG Ling.A plagiarism detection approach for Chinese documents based on semantic textual similarity[J].Journal of Shenzhen University Science and Engineering,2020,37(3):107.[doi:10.3724/SP.J.1249.2020.99107]
[6]宋昱,等.基于图像块l0梯度最小化的边缘保持平滑算法[J].深圳大学学报理工版,2021,38(3):307.[doi:10.3724/SP.J.1249.2021.03307]
 SONG Yu,and SUN Wenyun,et al.Edge-preserving smoothing algorithm based on l0 gradient minimization of image-patch[J].Journal of Shenzhen University Science and Engineering,2021,38(3):307.[doi:10.3724/SP.J.1249.2021.03307]

更新日期/Last Update: 2022-05-30