对抗长短时记忆网络的跨语言文本情感分类方法
|
摘要
针对文本情感分类任务中,有情感标注的语料在不同语言中的不均衡问题,结合深度学习和迁移学习,提出一种基于对抗长短时记忆网络(ALSTM)的跨语言文本情感分类方法.设置双语各自独立的特征提取网络和共享特征提取网络,把获取到的特征拼接输入到分类器进行分类.在共享特征提取网络中,设置语言分类器,运用对抗思想优化模型,通过投票法决定文本最终的情感极性.实验表明:该方法可以取得跨语言文本情感分类任务更高的准确度.
This paper proposes a cross-lingual sentiment classification method based on adversarial long short term memory(ALSTM)network,which aims at the problem of text sentiment classification in the disparity of emotionally annotated corpus in different languages,combined with deep learning and transfer learning.Bilingual feature extraction networks and a shared feature extraction network are set up,and then the extracted features are merged for classification.In the shared feature extraction network,a language classifier is set up.Using the adversarial idea to optimize the model,and the final polarity of the text depends on the voting results.Experiments show that cross-lingual sentiment classification can achieve higher accuracy by this method.
This paper proposes a cross-lingual sentiment classification method based on adversarial long short term memory(ALSTM)network,which aims at the problem of text sentiment classification in the disparity of emotionally annotated corpus in different languages,combined with deep learning and transfer learning.Bilingual feature extraction networks and a shared feature extraction network are set up,and then the extracted features are merged for classification.In the shared feature extraction network,a language classifier is set up.Using the adversarial idea to optimize the model,and the final polarity of the text depends on the voting results.Experiments show that cross-lingual sentiment classification can achieve higher accuracy by this method.
引文
[1]NAKOV P,RITTER A,ROSENTHAL S,et al.SemEval-2016task 4:Sentiment analysis in twitter[C]∥International Workshop on Semantic Evaluation.San Diego:[s.n.],2016:1-18.DOI:10.18653/v1/S16-1001.
[2]WAN Xiaojun.Using bilingual knowledge and ensemble techniques for unsupervised Chinese sentiment analysis[C]∥Conference on Empirical Methods in Natural Language Processing.Hawaii:DBLP,2008:553-561.DOI:10.3115/1613715.1613783.
[3]WAN Xiaojun.Co-training for cross-lingual sentiment classification[C]∥Proceedings of the Joint Conference of the47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP.Singapore:DBLP,2009:235-243.DOI:10.3115/1687878.1687913.
[4]LU Bin,TAN Chenhao,CARDIE C,et al.Joint bilingual sentiment classification with unlabeled parallel corpora[C]∥Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics:Human Language Technologies.Oregon:DBLP,2011:320-330.
[5]TANG Xuewei,WAN Xiaojun.Learning bilingual embedding model for cross-language sentiment classification[C]∥IEEE/WIC/ACM International Joint Conferences on Web Intelligence(WI)and Intelligent Agent Technologies(IAT).Warsaw:IEEE Press,2014:134-141.DOI:10.1109/WI-IAT.2014.90.
[6]方圆.跨语言文本情感分类技术研究[D].厦门:华侨大学,2015.
[7]ZHOU Huiwei,CHEN Long,SHI Fulin,et al.Learning bilingual sentiment word embeddings for cross-language sentiment classification[C]∥Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing.Beijing:[s.n.],2015:430-440.DOI:10.3115/v1/P15-1042.
[8]ZHOU Huiwei,YANG Yunlong,LIU Zhuang,et al.Jointly learning bilingual sentiment and semantic representations for cross-language sentiment classification[C]∥China Conference on Information Retrieval.Shanghai:Springer,2017:149-160.DOI:10.1007/978-3-319-68699-8_12.
[9]ZHOU Xinjie,WAN Xiaojun,XIAO Jiangguo.Attention-based LSTM network for cross-lingual sentiment classification[C]∥Conference on Empirical Methods in Natural Language Processing.Austin:Association for Computational Linguistics,2016:247-256.DOI:10.18653/v1/D16-1024.
[10]BEN-DAVID S,BLITZER J,CRAMMER K,et al.Analysis of representations for domain adaptation[C]∥International Conference on Neural Information Processing Systems.Vancouver:DBLP,2006:137-144.
[11]GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets[C]∥Advances in Neural Information Processing Systems.Montreal:MIT Press,2014:2672-2680.
[12]GANIN Y,LEMPITSKY V.Unsupervised domain adaptation by backpropagation[C]∥Proceedings of the 32nd International Conference on Machine Learning.Lille:arXiv,2014:1180-1189.
[13]PENNINGTON J,SOCHER R,MANNING C.GloVe:Global vectors for word representation[C]∥Conference on Empiricial Methods in Natural Language Processing.Doha:[s.n.],2014:1532-1543.DOI:10.3115/v1/D14-1162.
[14]TAI K S,SOCHER R,MANNING C D.Improved semantic representations from tree-structured long short-term memory networks[J].Computer Science,2015,5(1):36.DOI:10.3115/v1/P15-1150.
[15]IYER A,NATH S,SARAWAGI S.Maximum mean discrepancy for class ratio estimation:Convergence bounds and kernel selection[C]∥Proceedings of the 31st International Conference on Machine Learning.Beijing:JMLRorg,2014:1-530.
[16]GANIN Y,USTINOVA E,AJAKAN H,et al.Domain-adversarial training of neural networks[J].Journal of Machine Learning Research,2016,17(1):2096-2030.DOI:10.1007/978-3-319-58347-1_10.
[2]WAN Xiaojun.Using bilingual knowledge and ensemble techniques for unsupervised Chinese sentiment analysis[C]∥Conference on Empirical Methods in Natural Language Processing.Hawaii:DBLP,2008:553-561.DOI:10.3115/1613715.1613783.
[3]WAN Xiaojun.Co-training for cross-lingual sentiment classification[C]∥Proceedings of the Joint Conference of the47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP.Singapore:DBLP,2009:235-243.DOI:10.3115/1687878.1687913.
[4]LU Bin,TAN Chenhao,CARDIE C,et al.Joint bilingual sentiment classification with unlabeled parallel corpora[C]∥Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics:Human Language Technologies.Oregon:DBLP,2011:320-330.
[5]TANG Xuewei,WAN Xiaojun.Learning bilingual embedding model for cross-language sentiment classification[C]∥IEEE/WIC/ACM International Joint Conferences on Web Intelligence(WI)and Intelligent Agent Technologies(IAT).Warsaw:IEEE Press,2014:134-141.DOI:10.1109/WI-IAT.2014.90.
[6]方圆.跨语言文本情感分类技术研究[D].厦门:华侨大学,2015.
[7]ZHOU Huiwei,CHEN Long,SHI Fulin,et al.Learning bilingual sentiment word embeddings for cross-language sentiment classification[C]∥Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing.Beijing:[s.n.],2015:430-440.DOI:10.3115/v1/P15-1042.
[8]ZHOU Huiwei,YANG Yunlong,LIU Zhuang,et al.Jointly learning bilingual sentiment and semantic representations for cross-language sentiment classification[C]∥China Conference on Information Retrieval.Shanghai:Springer,2017:149-160.DOI:10.1007/978-3-319-68699-8_12.
[9]ZHOU Xinjie,WAN Xiaojun,XIAO Jiangguo.Attention-based LSTM network for cross-lingual sentiment classification[C]∥Conference on Empirical Methods in Natural Language Processing.Austin:Association for Computational Linguistics,2016:247-256.DOI:10.18653/v1/D16-1024.
[10]BEN-DAVID S,BLITZER J,CRAMMER K,et al.Analysis of representations for domain adaptation[C]∥International Conference on Neural Information Processing Systems.Vancouver:DBLP,2006:137-144.
[11]GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets[C]∥Advances in Neural Information Processing Systems.Montreal:MIT Press,2014:2672-2680.
[12]GANIN Y,LEMPITSKY V.Unsupervised domain adaptation by backpropagation[C]∥Proceedings of the 32nd International Conference on Machine Learning.Lille:arXiv,2014:1180-1189.
[13]PENNINGTON J,SOCHER R,MANNING C.GloVe:Global vectors for word representation[C]∥Conference on Empiricial Methods in Natural Language Processing.Doha:[s.n.],2014:1532-1543.DOI:10.3115/v1/D14-1162.
[14]TAI K S,SOCHER R,MANNING C D.Improved semantic representations from tree-structured long short-term memory networks[J].Computer Science,2015,5(1):36.DOI:10.3115/v1/P15-1150.
[15]IYER A,NATH S,SARAWAGI S.Maximum mean discrepancy for class ratio estimation:Convergence bounds and kernel selection[C]∥Proceedings of the 31st International Conference on Machine Learning.Beijing:JMLRorg,2014:1-530.
[16]GANIN Y,USTINOVA E,AJAKAN H,et al.Domain-adversarial training of neural networks[J].Journal of Machine Learning Research,2016,17(1):2096-2030.DOI:10.1007/978-3-319-58347-1_10.