Please wait a minute...
Big Data Mining and Analytics  2021, Vol. 4 Issue (3): 195-207    DOI: 10.26599/BDMA.2021.9020003
    
A Multitask Multiview Neural Network for End-to-End Aspect-Based Sentiment Analysis
Yong Bie1(),Yan Yang1,*()
School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu 611756, China
Download: PDF (614 KB)      HTML  
Export: BibTeX | EndNote (RIS)      

Abstract  

The aspect-based sentiment analysis (ABSA) consists of two subtasks'aspect term extraction and aspect sentiment prediction. Existing methods deal with both subtasks one by one in a pipeline manner, in which there lies some problems in performance and real application. This study investigates the end-to-end ABSA and proposes a novel multitask multiview network (MTMVN) architecture. Specifically, the architecture takes the unified ABSA as the main task with the two subtasks as auxiliary tasks. Meanwhile, the representation obtained from the branch network of the main task is regarded as the global view, whereas the representations of the two subtasks are considered two local views with different emphases. Through multitask learning, the main task can be facilitated by additional accurate aspect boundary information and sentiment polarity information. By enhancing the correlations between the views under the idea of multiview learning, the representation of the global view can be optimized to improve the overall performance of the model. The experimental results on three benchmark datasets show that the proposed method exceeds the existing pipeline methods and end-to-end methods, proving the superiority of our MTMVN architecture.



Key wordsdeep learning      multitask learning      multiview learning      natural language processing      aspect-based sentiment analysis     
Received: 03 January 2021      Published: 20 May 2021
Fund:  National Natural Science Foundation of China(61976247)
Corresponding Authors: Yan Yang     E-mail: autwind_by@163.com;yyang@swjtu.edu.cn
About author: Yong Bie received the BS degree from Southwest Jiaotong University, Chengdu, China in 2018 and now he is a master candidate of Southwest Jiaotong University. His current research interests include data mining, natural language processing, and machine learning.|Yan Yang received the BS and MS degrees from Huazhong University of Science and Technology, Wuhan, China in 1984 and 1987, respectively. She received the PhD degree from Southwest Jiaotong University, Chengdu, China in 2007. From 2002 to 2003 and 2004 to 2005, she was a visiting scholar with the University of Waterloo, Waterloo, Canada. She is currently a professor and vice dean at the School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu, China. Her current research interests include multi-view learning, big data analysis and mining, ensemble learning, semi-supervised learning, and cloud computing.
Cite this article:

Yong Bie,Yan Yang. A Multitask Multiview Neural Network for End-to-End Aspect-Based Sentiment Analysis. Big Data Mining and Analytics, 2021, 4(3): 195-207.

URL:

http://bigdata.tsinghuajournals.com/10.26599/BDMA.2021.9020003     OR     http://bigdata.tsinghuajournals.com/Y2021/V4/I3/195

Fig. 1 Architecture of proposed MTMVN model.
Gold labelsInput sequence
Mouleswereexcellent,lobsterravioliwasverysalty!
Gold labels for AEBOOOBIOOOO
Gold labels for ASPPOS---NEGNEG----
Table 1 A labeling example with gold AE and ASP labels.
DatasetTrainTest
SentenceAspect termsSentenceAspect terms
DL30452358800654
DR304136938001134
DT21152608235604
Table 2 Numbers of sentences and aspect terms contained by three benchmark datasets.
DatasetTrainTestTotal
DLPOS9873411328
NEG866128994
NEU460169629
CONF451661
DRPOS21647282892
NEG8051961001
NEU633196829
CONF9114105
DTPOS563132695
NEG21848266
NEU18274242251
Table 3 Numbers of aspects holding different sentiment polarities in three benchmark datasets.
MethodModelF1 scores (%)
DLDRDT
Pipeline methodsCRF-pipeline51.7354.1631.35
NN-CRF-pipeline53.2361.0345.08
CMLA-ALSTM53.6863.8746.47
DECNN-ALSTM54.8765.1347.24
End-to-End MethodsSentiment-Scope50.2762.0145.91
LSTM-CRF-LSTMc54.3765.0346.82
LSTM-CRF-CNNc54.7164.2947.35
MNN53.9363.7146.16
OursMTMVN55.0865.2047.89
Table 4 Comparison results in the F1 scores (%) of all the baselines and our proposed model for the whole ABSA task.
Model VariantsF1 scores (%)
DLDRDT
Only UFT52.3259.3544.85
Only Fusion52.7361.6944.20
Fusion+AE+ASP53.4862.3645.64
UFT+Fusion+AE+ASP54.1263.7746.78
The entire MTMVN model55.0865.2047.89
Table 5 Comparison results in the F1 scores (%) of different model variants.
DL with different σ.
">
Fig. 2 F1 score (%) on DL with different σ.
DR with different σ.
">
Fig. 3 F1 score (%) on DR with different σ.
ExampleModel
CMLA-ALSTMMNNOur MTMVN
No [installation disk (dvd)]NEU is included.[installation disk]NEU (?)[installation disk]NEG, [dvd]NEG (?)[installation disk]NEU, [dvd]NEU (?)
[Works]POS well, and I am extremely happy to be back to an [apple os]POS.[Works]POS (?),
[apple os]POS (?)
None (?)[Works]POS (?),
[apple os]POS (?)
Straight-forward, no surprises, very decent [japanese food]POS.[japanese food]NEG (?)None (?)[japanese food]POS (?)
Try the [rose roll]POS (not on [menu]NEU).[rose roll]POS (?),
[menu]POS (?)
[rose roll]POS (?),
None (?)
[rose roll]POS (?),
[menu]NEU (?)
The only task that this computer would not be good enough for be [gaming]NEG, otherwise the [integrated Intel 4000 graphics]CONF work well for other tasks.[gaming]NEG (?),
[graphics]POS, [tasks]NEG (?)
[gaming]NEG (?),
[graphics]POS (?)
[gaming]NEG (?),
[integrated Intel graphics]NEG (?)
Table 6 Case analysis on CMLA_ALSTM, MNN, and MTMVN; ? and ? mean wrong and correct predictions, respectively.
[1]   Bouazizi M., Ohtsuki T., Multi-class sentiment analysis on twitter: Classification performance and challenges, Big Data Mining and Analytics, vol. 2, no. 3, pp. 181-194, 2019.
[2]   Liu B., Sentiment Analysis and Opinion Mining. San Rafael, CA, USA: Morgan & Claypool Publishers, 2012.
[3]   Jiang P., Zhang C. X., Fu H. P., Niu Z. D., and Yang Q., An approach based on tree kernels for opinion mining of online product reviews, in 2010 IEEE Int. Conf. Data Mining, Sydney, Australia, 2010, pp. 256-265.
[4]   Jin W. and Ho H. H., A novel lexicalized hmm-based learning framework for web opinion mining, in Proc. 26th Int. Conf. Machine Learning, Montreal, Canada, 2009, pp. 465-472.
[5]   Jakob N. and Gurevych I., Extracting opinion targets in a single- and cross-domain setting with conditional random fields, in Proc. 2010 Conf. Empirical Methods in Natural Language Proc., Boston, MA, USA, 2010, pp. 1035-1045.
[6]   Li F. T., Han C., Huang M. L., Zhu X. Y., Xia Y. J., Zhang S., and Yu H., Structure-aware review mining and summarization, in Proc. 23rd Int. Conf. Computational Linguistics, Beijing, China, 2010, pp. 653-661.
[7]   Poria S., Cambria E., and Gelbukh A., Aspect extraction for opinion mining with a deep convolutional neural network, Knowl. Based Syst., vol. 108, pp. 42-49, 2016.
[8]   Wang W. Y., Pan S. J., Dahlmeier D., and Xiao X. K., Coupled multi-layer attentions for co-extraction of aspect and opinion terms, in Proc. 31st AAAI Conf. Artificial Intelligence, San Francisco, CA, USA, 2017, pp. 3316-3322.
[9]   Ye H., Yan Z. C., Luo Z. C., and Chao W. H., Dependency-tree based convolutional neural networks for aspect term extraction, in Pacific-Asia Conf. Knowledge Discovery and Data Mining, H. Ye, Z. Yan, Z. Luo, and W. Chao, Eds. Cham, Germany: Springer, 2017, pp. 350-362.
[10]   Luo H. S., Li T. R., Liu B., Wang B., and Unger H., Improving aspect term extraction with bidirectional dependency tree representation, IEEE/ACM Trans. Audio Speech Language Proc., vol. 27, no. 7, pp. 1201-1212, 2019.
[11]   Xu H., Liu B., Shu L., and Yu P. S., Double embeddings and CNN-based sequence labeling for aspect extraction, in Proc. 56th Ann. Meeting of the Association for Computational Linguistics, Melbourne, Australia, 2018, pp. 592-598.
[12]   Wang Y. Q., Huang M. L., Zhu X. Y., and Zhao L., Attention-based LSTM for aspect-level sentiment classification, in Proc. 2016 Conf. Empirical Methods in Natural Language Proc., Austin, TX, USA, 2016, pp. 606-615.
[13]   Tang D. Y., Qin B., and Liu T., Aspect level sentiment classification with deep memory network, in Proc. 2016 Conf. Empirical Methods in Natural Language Proc., Austin, TX, USA, 2016, pp. 214-224.
[14]   Fan F. F., Feng Y. S., and Zhao D. Y., Multi-grained attention network for aspect-Level sentiment classification, in Proc. 2018 Conf. Empirical Methods in Natural Language Proc., Brussels, Belgium, 2018, pp. 3433-3442.
[15]   Chen P., Sun Z. Q., Bing L. D., and Yang W., Recurrent attention network on memory for aspect sentiment analysis, in Proc. 2017 Conf. Empirical Methods in Natural Language Proc., Copenhagen, Denmark, 2017, pp. 452-461.
[16]   Ma D. H., Li S. J., Zhang X. D., and Wang H. F., Interactive attention networks for aspect-level sentiment classification, in Proc. 26th Int. Joint Conf. Artificial Intelligence, Melbourne, Australia, 2017, pp. 4068-4074.
[17]   Huang B. X., Ou Y. L., and Carley K. M., Aspect level sentiment classification with attention-over-attention neural networks, in Int. Conf. Social Computing, Behavioral-Cultural Modeling & Prediction and Behavior Representation in Modeling and Simulation (SBP-BRiMS), Thomson R., Dancy C., Hyder A., and Bisgin H., Eds. Cham, Germany: Springer, 2018, pp. 197-206.
[18]   Li L. S., Liu Y., and Zhou A. Q., Hierarchical attention based position-aware network for aspect-level sentiment analysis, in Proc. 22nd Conf. Computational Natural Language Learning, Brussels, Belgium, 2018, pp. 181-189.
[19]   Song Y. W., Wang J. H., Jiang T., Liu Z. Y., and Rao Y. H., Attentional encoder network for targeted sentiment classification, arXiv preprint arXiv: 1902.09314v2, 2019.
[20]   Li X., Bing L. D., Lam W., and Shi B., Transformation networks for target-oriented sentiment classification, in Proc. 56th Ann. Meeting of the Association for Computational Linguistics, Melbourne, Australia, 2018, pp. 946-956.
[21]   Wang F. X., Lan M., and Wang W. T., Towards a one-stop solution to both aspect extraction and sentiment analysis tasks with neural multi-task learning, in Proc. 2018 Int. Joint Conf. Neural Networks (IJCNN), Rio de Janeiro, Brazil, 2018, pp. 1-8.
[22]   Li X., Bing L. D., Li P. J., and Lam W., A unified model for opinion target extraction and target sentiment prediction, Proc. AAAI Conf. Artif. Intell., vol. 33, no. 1, pp. 6714-6721, 2019.
[23]   He R. D., Lee W. S., Ng H. T., and Dahlmeier D., An interactive multi-task learning network for end-to-end aspect-based sentiment analysis, in Proc. 57th Ann. Meeting of the Association for Computational Linguistics, Florence, Italy, 2019, pp. 504-515.
[24]   Luo H. S., Li T. R., Liu B., and Zhang J. B., DOER: Dual cross-shared RNN for aspect term-polarity co-extraction, in Proc. 57th Ann. Meeting of the Association for Computational Linguistics, Florence, Italy, 2019, pp. 591-601.
[25]   Kim Y., Convolutional neural networks for sentence classification, in Proc. 2014 Conf. Empirical Methods in Natural Language Proc., Doha, Qatar, 2014, pp. 1746-1751.
[26]   Cho K., Van Merri?nboer B., Gulcehre C., Bahdanau D., Bougares F., Schwenk H., and Bengio Y., Learning phrase representations using RNN encoder-decoder for statistical machine translation, in Proc. 2014 Conf. Empirical Methods in Natural Language Processing, Doha, Qatar, 2014, pp. 1724-1734.
[27]   Sha L., Zhang X. D., Qian F., Chang B. B., and Sui Z. F., A Multi-view fusion neural network for answer selection, in 32nd AAAI Conf. Artificial Intelligence, New Orleans, LA, USA, 2018, pp. 5422-5429.
[28]   Andrew G., Arora R., Bilmes J., and Livescu K., Deep canonical correlation analysis, in Proc. 30th Int. Conf. Machine Learning, Atlanta, GA, USA, 2013, pp. 1247-1255.
[29]   Hochreiter S., Schmidhuber J., Long short-term memory, Neural Comput., vol. 9, no. 8, pp. 1735-1780, 1997.
[30]   Weston J., Chopra S., and Bordes A., Memory networks, arXiv preprint arXiv: 1410.3916, 2014.
[31]   Cui Y. M., Chen Z. P., Wei S., Wang S. J., Liu T., and Hu G. P., Attention-over-attention neural networks for reading comprehension, in Proc. 55th Ann. Meeting of the Association for Computational Linguistics, Vancouver, Canada, 2017, pp. 593-602.
[32]   Pennington J., Socher R., and Manning C. D., Glove: Global vectors for word representation, in Proc. 2014 Conf. Empirical Methods in Natural Language Proc., Doha, Qatar, 2014, pp. 1532-1543.
[33]   He K. M., Zhang X. Y., Ren S. Q., and Sun J., Deep residual learning for image recognition, in Proc. 2016 IEEE Conf. Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 2016, pp. 770-778.
[34]   Pontiki M., Galanis D., Pavlopoulos J., Papageorgiou H., Androutsopoulos I., and Manandhar S., Semeval-2014 task 4: Aspect based sentiment analysis, in Proc. 8th Int. Workshop on Semantic Evaluation (SemEval 2014), Dublin, Ireland, 2014, pp. 27-35.
[35]   Mitchell M., Aguilar J., Wilson T., and Van Durme B., Open domain targeted sentiment, in Proc. 2013 Conf. Empirical Methods in Natural Language Proc., Seattle, WA, USA, 2013, pp. 1643-1654.
[36]   Zhang M. S., Zhang Y., and Vo D. T., Neural networks for open domain targeted sentiment, in Proc. 2015 Conf. Empirical Methods in Natural Language Proc., Lisbon, Portugal, 2015, pp. 612-621.
[37]   Li H. and Lu W., Learning latent sentiment scopes for entity-level sentiment analysis, in Proc. 31st AAAI Conf. Artificial Intelligence, San Francisco, CA, USA, 2017, pp. 3482-3489.
[38]   Lample G., Ballesteros M., Subramanian S., Kawakami K., and Dyer C., Neural architectures for named entity recognition, in Proc. 2016 Conf. North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego, CA, USA, 2016, pp. 260-270.
[39]   Ma X. Z. and Hovy E., End-to-end sequence labeling via bi-directional LSTM-CNNs-CRF, in Proc. 54th Annu. Meeting of the Association for Computational Linguistics, Berlin, Germany, 2016, pp. 1064-1074.
[40]   Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A. N., Kaiser ?, and Polosukhin I., Attention is all you need, in Proc. 31st Conf. Neural Information Proc. Systems, Long Beach, CA, USA, 2017, pp. 5998-6008.
[41]   Glorot X. and Bengio Y., Understanding the difficulty of training deep feedforward neural networks, in Proc. 13th Int. Conf. Artificial Intelligence and Statistics, Sardinia, Italy, vol. 9, pp. 249-256.
[42]   He K. M., Zhang X. Y., Ren S. Q., and Sun J., Delving deep into rectifiers: surpassing human-level performance on ImageNet classification, in 2015 IEEE Int. Conf. Computer Vision (ICCV), Santiago, Chile, 2015, pp. 1026-1034.
[43]   Scarselli F., Gori M., Tsoi A. C., Hagenbuchner M., and Monfardini G., The graph neural network model, IEEE Trans. Neural Networks, vol. 20, no. 1, pp. 61-80, 2009.
[1] Shuai Zhang,Hongyan Liu,Jun He,Sanpu Han,Xiaoyong Du. Deep Sequential Model for Anchor Recommendation on Live Streaming Platforms[J]. Big Data Mining and Analytics, 2021, 4(3): 173-182.
[2] Krishna Kant Singh,Akansha Singh. Diagnosis of COVID-19 from Chest X-Ray Images Using Wavelets-Based Depthwise Convolution Network[J]. Big Data Mining and Analytics, 2021, 4(2): 84-93.
[3] Natarajan Yuvaraj,Kannan Srihari,Selvaraj Chandragandhi,Rajan Arshath Raja,Gaurav Dhiman,Amandeep Kaur. Analysis of Protein-Ligand Interactions of SARS-CoV-2 Against Selective Drug Using Deep Neural Networks[J]. Big Data Mining and Analytics, 2021, 4(2): 76-83.
[4] Youssef Nait Malek,Mehdi Najib,Mohamed Bakhouya,Mohammed Essaaidi. Multivariate Deep Learning Approach for Electric Vehicle Speed Forecasting[J]. Big Data Mining and Analytics, 2021, 4(1): 56-64.
[5] Wei Zhong, Ning Yu, Chunyu Ai. Applying Big Data Based Deep Learning System to Intrusion Detection[J]. Big Data Mining and Analytics, 2020, 3(3): 181-195.
[6] Sunitha Basodi, Chunyan Ji, Haiping Zhang, Yi Pan. Gradient Amplification: An Efficient Way to Train Deep Neural Networks[J]. Big Data Mining and Analytics, 2020, 3(3): 196-207.
[7] Chaity Banerjee, Tathagata Mukherjee, Eduardo Pasiliao Jr.. Feature Representations Using the Reflected Rectified Linear Unit (RReLU) Activation[J]. Big Data Mining and Analytics, 2020, 3(2): 102-120.
[8] Lujia Shen, Qianjun Liu, Gong Chen, Shouling Ji. Text-Based Price Recommendation System for Online Rental Houses[J]. Big Data Mining and Analytics, 2020, 3(2): 143-152.
[9] Zhenxing Guo, Shihua Zhang. Sparse Deep Nonnegative Matrix Factorization[J]. Big Data Mining and Analytics, 2020, 03(01): 13-28.
[10] Qile Zhu, Xiyao Ma, Xiaolin Li. Statistical Learning for Semantic Parsing: A Survey[J]. Big Data Mining and Analytics, 2019, 2(4): 217-239.
[11] Ying Yu, Min Li, Liangliang Liu, Yaohang Li, Jianxin Wang. Clinical Big Data and Deep Learning: Applications, Challenges, and Future Outlooks[J]. Big Data Mining and Analytics, 2019, 2(4): 288-305.
[12] Wenmao Wu, Zhizhou Yu, Jieyue He. A Semi-Supervised Deep Network Embedding Approach Based on the Neighborhood Structure[J]. Big Data Mining and Analytics, 2019, 2(3): 205-216.
[13] Jiangcheng Zhu, Shuang Hu, Rossella Arcucci, Chao Xu, Jihong Zhu, Yi-ke Guo. Model Error Correction in Data Assimilation by Integrating Neural Networks[J]. Big Data Mining and Analytics, 2019, 2(2): 83-91.
[14] Jin Liu, Yi Pan, Min Li, Ziyue Chen, Lu Tang, Chengqian Lu, Jianxin Wang. Applications of Deep Learning to MRI Images: A Survey[J]. Big Data Mining and Analytics, 2018, 1(1): 1-18.
[15] Ning Yu, Zhihua Li, Zeng Yu. Survey on Encoding Schemes for Genomic Data Representation and Feature Learning—From Signal Processing to Machine Learning[J]. Big Data Mining and Analytics, 2018, 01(03): 191-210.