Max-Cosine Matching Based Neural Models for Recognizing Textual Entailment

被引:1
作者
Xie, Zhipeng [1 ]
Hu, Junfeng [1 ]
机构
[1] Fudan Univ, Shanghai Key Lab Data Sci, Sch Comp Sci, Shanghai, Peoples R China
来源
DATABASE SYSTEMS FOR ADVANCED APPLICATIONS (DASFAA 2017), PT I | 2017年 / 10177卷
关键词
Textual entailment; Recurrent neural networks; LSTM; ACQUISITION;
D O I
10.1007/978-3-319-55753-3_19
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Recognizing textual entailment is a fundamental task in a variety of text mining or natural language processing applications. This paper proposes a simple neural model for RTE problem. It first matches each word in the hypothesis with its most-similar word in the premise, producing an augmented representation of the hypothesis conditioned on the premise as a sequence of word pairs. The LSTM model is then used to model this augmented sequence, and the final output from the LSTM is fed into a softmax layer to make the prediction. Besides the base model, in order to enhance its performance, we also proposed three techniques: the integration of multiple word-embedding library, bi-way integration, and ensemble based on model averaging. Experimental results on the SNLI dataset have shown that the three techniques are effective in boosting the predicative accuracy and that our method outperforms several state-of-the-state ones.
引用
收藏
页码:295 / 308
页数:14
相关论文
共 30 条
[1]  
Aharon R., 2010, P 48 ANN M ASS COMPU, P241
[2]  
[Anonymous], 2016, ARXIV150906664
[3]  
[Anonymous], 1998, WordNet, DOI DOI 10.7551/MITPRESS/7287.001.0001
[4]  
[Anonymous], 2005, P IWP
[5]  
Baker C.F., 1998, P 36 ANN M ASS COMP, P86, DOI DOI 10.3115/980845.980860
[6]   LEARNING LONG-TERM DEPENDENCIES WITH GRADIENT DESCENT IS DIFFICULT [J].
BENGIO, Y ;
SIMARD, P ;
FRASCONI, P .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1994, 5 (02) :157-166
[7]  
Bowman S.R., 2015, EMNLP, P632, DOI 10.18653/v1/D15
[8]  
Bowman SR, 2016, PROCEEDINGS OF THE 54TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, VOL 1, P1466
[9]  
Chen JZ, 2016, PROCEEDINGS OF 2016 12TH INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE AND SECURITY (CIS), P551, DOI [10.1109/CIS.2016.133, 10.1109/CIS.2016.0134]
[10]   Recognizing textual entailment: Rational, evaluation and approaches [J].
Dagan, Ido ;
Dolan, Bill ;
Magnini, Bernardo ;
Roth, Dan .
Natural Language Engineering, 2009, 15 (04) :i-xvii