第2期 刘杰,等:基于双语对抗学习的半监督情感分类 63 误标签的问题。同时本文使用了中文语料对应的英文语料的对抗学习,通过中英文对抗神经网络的联合学 习有效提高了半监督情感分类任务的性能 参考文献: [1] PANG B, LEE L, VAITHYANATHAN S. Sentiment classification using machine learning techniques [c]//Proceedings of the Empirical Methods in Natural Language Processing. Philadelphia, 2002: 79-86 [2] ZAGIBALOV T, CARROLL J. Automatie seed word selection for unsupervised sentiment classification of Chinese text [c]// Proceedings of the: International Conference on Computational Linguistics. Manchester, 2008: 1073-1080 3] LIN C H, HE Y L. Joint sentiment/topic model for sentiment analysis [c]// Proceeding of the 18th ACM Conference on Infor- mation& Knowledge Management. Hong Kong, 2009: 375-384. [4 WAN X J. Co-training for cross-lingual sentiment classification [ c]// Proceedings of Joint Conference of the Meeting of the ACI and the International Joint Conference on Natural Language. Singapore, 2009: 235-243 [5 ZHU X J, GHAHRAMANI Z. Learning from labeled and unlabeled data with label propagation [J]. Tech Report, 2002, 3175 (2004):237-244. [6] ZHOU SS, CHEN Q C, WANG X L. Active deep networks for semi-supervised sentiment elassification [c]//Proceedings of he International Conference on Computational Linguistics. Beijing, 2010: 1515-1523 [7]王志昊.情感分类特征选择方法[D].苏州:苏州大学,2014 WANG Z H. Research on feature selection for sentiment classification [D]. Suzhou: Soochow University, 2014. [8]苏艳.双语情感分类方法研究[D].苏州:苏州大学,2013 SU Y. Sentiment elassification with bilingual text [D]- Suzhou: Soochow University, 2013 [9] KINGMA D P, BA J. Adam: a method for stochastic optimization C]//International Conference on Learning Representions 2015:1-13 [10 WAN X J. Bilingual co-training for sentiment classification of Chinese produet reviews[ J]. Computational linguistics, 2011, 3 (3):587-616 [11] XU W D, SUN H Z, DENG C, et al. Variational autoencoders for semi-supervised text classification [c]//Proceedings of the 31st AAAl Conference on Artificial Intelligence. San Francisco. 2017: 3358-3364 Semi-supervised Sentiment Classification with Bil Adversarial Learning Liu Jie, LIU Huan, LI Shoushan YAN Wei (1. Institute of Information Engineering, Suqian College, Suqian 223800, China; 2. School of Computer Science Technology, Soochon University, Suzhou 215006, China) Abstract: A bilingual adversarial learning approach was proposed to make full use of the information of abeled samples. Specifically, the labeled and unlabeled Chinese samples were encoded by independ ent LSTMs. and then fed into classifier and discriminator. The function of classifier was to make the la- beled samples and unlabeled in the same distribution, while the discriminator was used to distinguish whether the input sample was labeled and unlabeled. Finally, another adversarial neural network with the English samples was constructed, and the performance of semi-supervised sentiment classification was ex lish adversarial networks studies showed that the proposed approach achieved good accuracy on different sizes of training sets, and demonstrated the significant improvement compared to other baselines. Key words: unlabeled samples; bilingual adversarial learning; semi-supervised sentiment classification (责任编辑:王浩毅)第 2 期 刘 杰,等:基于双语对抗学习的半监督情感分类 误标签的问题。 同时本文使用了中文语料对应的英文语料的对抗学习,通过中英文对抗神经网络的联合学 习有效提高了半监督情感分类任务的性能。 参考文献: [1] PANG B, LEE L, VAITHYANATHAN S. Sentiment classification using machine learning techniques [C]∥ Proceedings of the Empirical Methods in Natural Language Processing. Philadelphia, 2002: 79-86. [2] ZAGIBALOV T, CARROLL J. Automatic seed word selection for unsupervised sentiment classification of Chinese text [ C]∥ Proceedings of the International Conference on Computational Linguistics. Manchester, 2008: 1073-1080. [3] LIN C H, HE Y L. Joint sentiment / topic model for sentiment analysis [C]∥ Proceeding of the 18th ACM Conference on Infor￾mation & Knowledge Management. Hong Kong, 2009: 375-384. [4] WAN X J. Co-training for cross-lingual sentiment classification [C]∥ Proceedings of Joint Conference of the Meeting of the ACL and the International Joint Conference on Natural Language. Singapore, 2009: 235-243. [5] ZHU X J, GHAHRAMANI Z. Learning from labeled and unlabeled data with label propagation [ J] . Tech Report, 2002, 3175 (2004) : 237-244. [6] ZHOU S S, CHEN Q C, WANG X L. Active deep networks for semi-supervised sentiment classification [ C]∥ Proceedings of the International Conference on Computational Linguistics. Beijing, 2010: 1515-1523. [7] 王志昊. 情感分类特征选择方法[D] . 苏州:苏州大学,2014. WANG Z H. Research on feature selection for sentiment classification [D] . Suzhou: Soochow University, 2014. [8] 苏艳. 双语情感分类方法研究[D] . 苏州:苏州大学,2013. SU Y. Sentiment classification with bilingual text [D] . Suzhou: Soochow University, 2013. [9] KINGMA D P, BA J. Adam: a method for stochastic optimization [ C]∥International Conference on Learning Representions. San Diego, 2015:1-13. [10] WAN X J. Bilingual co-training for sentiment classification of Chinese product reviews[ J] . Computational linguistics, 2011, 37 (3) : 587-616. [11] XU W D, SUN H Z, DENG C, et al. Variational autoencoders for semi-supervised text classification [C]∥ Proceedings of the 31st AAAI Conference on Artificial Intelligence. San Francisco, 2017: 3358-3364. Semi-supervised Sentiment Classification with Bilingual Adversarial Learning LIU Jie 1 , LIU Huan 2 , LI Shoushan 2 , YAN Wei 1 (1. Institute of Information Engineering, Suqian College, Suqian 223800, China; 2. School of Computer Science & Technology, Soochow University, Suzhou 215006, China) Abstract: A bilingual adversarial learning approach was proposed to make full use of the information of unlabeled samples. Specifically, the labeled and unlabeled Chinese samples were encoded by independ￾ent LSTMs, and then fed into classifier and discriminator. The function of classifier was to make the la￾beled samples and unlabeled in the same distribution, while the discriminator was used to distinguish whether the input sample was labeled and unlabeled. Finally, another adversarial neural network with the English samples was constructed, and the performance of semi-supervised sentiment classification was ex￾pected to be improved through the joint learning of Chinese and English adversarial networks. Empirical studies showed that the proposed approach achieved good accuracy on different sizes of training sets, and demonstrated the significant improvement compared to other baselines. Key words: unlabeled samples; bilingual adversarial learning; semi-supervised sentiment classification (责任编辑:王浩毅) 63