| 262 | 5 | 394 |
| 下载次数 | 被引频次 | 阅读次数 |
提出一个新的情感回归半监督领域适应方法.首先使用长短期记忆网络(long short-term memory,LSTM)实现回归模型,其次使用变分自编码器(variational autoencoder,VAE)实现生成模型,最后联合学习LSTM回归模型和VAE生成模型,实现基于变分自编码器的情感回归半监督领域适应模型.实验结果表明,所提出的基于变分自编码器的情感回归半监督领域适应方法较其他基准方法能有效提高实验性能.
Abstract：A novel approach was proposed for semi-supervised domain adaptation of sentiment regression,namely VAE-R. In VAE-R model,a long short-term memory( LSTM) network was employed to achieve a regression model,and then a generation model based on variational autoencoder( VAE). In the learning process,the LSTM regression model and the VAE generation model were jointly learned. Empirical studies had demonstrated the effectiveness of the proposed approach in the semi-supervised domain adaptation.
[1] PANG B,LEE L. Opinion mining and sentiment analysis[J]. Foundations and trendsin information retrieval,2008,2(1/2):1-135.
[2] LIU P,QIU X,HUANG X. Adversarial multi-task learning for text classification[C]∥Proceedings of the 56th Annual Meetingof the Association for Computational Linguistics. Vanciuver,2017:1-10.
[3] GEHRING J,AULI M,GRANGIER D,et al. A convolutional encoder model for neural machine translation[C]∥Proceedingsof the 56th Annual Meeting of the Association for Computational Linguistics. Vanciuver,2017:123-135.
[4] CUI Y,CHEN Z,WEI S,et al. Attention-over-attention neural networks for reading comprehension[EB/OL].(2017-06-06)[2018-07-01]. https:∥arxiv.org/pdf/1607. 04423.pdf.
[5] BLITZER J,DREDZE M,PEREIRA F. Biographies,bollywood,boom-boxes and blenders:domain adaptation for sentimentclassification[C]∥Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics. Prague,2007,31(2):187-205.
[6] GLOROT X,BORDES A,BENGIO Y. Domain adaptation for large-scale sentiment classification:a deep learning approach[C]∥International Conference on Machine Learning. Omnipress,2011:513-520.
[7] BENGIO Y,LAMBLIN P,POPOVICI D,et al. Greedy layer-wise training of deep networks[C]∥Advances in Neural Informa-tion Processing Systems. Vancouver,2007:153-160.
[8] KINGMA D,MOHAMED S,REZENDE D,et al. Semi-supervised learning with deep generative models[C]∥Advances inNeural Information Processing Systems. Montréal,2014:3581-3589.
[9] REZENDE D,MOHAMED S,WIERSTRA D. Stochastic backpropagation and approximate inference in deep generative models[C]∥Proceedings of International Conference on Machine Learning. Beijing,2014:1278-1286.
[10] XU W,SUN H,DENG C,et al. Variational autoencoder for semi-supervised text classification[C]∥Proceedings of the Ar-tifitial Intelligence. San Francisco,2017:781-790.
[11] WEN T,GASIC M,MRKSIC N,et al. Semantically conditioned LSTM-based natural language generation for spoken dialoguesystems[C]∥Proceedings of the Empirical Methods in Natural Language. Lisbon,2015:1711-1721.
[12] GHOSH S,VINYALS O,STROPE B,et al. Contextual LSTM(CLSTM)models for large scale nlp tasks[EB/OL].(2016-02-19)[2018-07-01]. http:∥cn.arxiv.org/pdf/1602. 06291.
[13] SERBAN I V,SORDONI A,BENGIO Y,et al. Building end-to-end dialogue systems using generative hierarchical neural net-work models[C]∥Thirtieth AAAI Conference on Artificial Intelligence. Phoenix,2016:3776-3783.
[14] MCAULEY J,PANDEY R,LESKOVEC J. Inferring networks of substitutable and complementary products[C]∥Proceedings ofthe 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Sydney,2015:785-794.
[15] CAMERON A,WINDMEIJER F. R-squared measures for count data regression models with applications to health care utiliza-tion[J]. Journal of business&economic statistics,1996,14(2):209-220.
基本信息:
DOI：10.13705/j.issn.1671-6841.2018216
中图分类号:TP391.1
引用信息:
[1]刘欢,徐健,李寿山.基于变分自编码器的情感回归半监督领域适应方法[J],2019,51(02):47-51.DOI:10.13705/j.issn.1671-6841.2018216.
基金信息:
国家自然科学基金项目(61331011,61375073)
2018-07-16
2018
2018-12-03
2018
1