Skip to main content
SpringerLink
Log in

Representation learning from noisy user-tagged data for sentiment classification

  • Original Article
  • Published:
International Journal of Machine Learning and Cybernetics Aims and scope Submit manuscript

Abstract

Sentiment classification aims to identify the sentiment orientation of an opinionated text, which is widely used for market research, product recommendation, and etc. Supervised deep learning approaches are prominent in sentiment classification and have shown the power in representation learning, however such methods suffer from the costly human annotations. Massive user-tagged opinionated texts on the Internet provide a new source for annotation, such as twitter with emoji. However, the texts may contain noisy labels, which may cause ambiguity during training process. In this paper, we propose a novel Weakly-supervised Anti-noise Contrastive Learning framework for sentiment classification, and name it as WACL. We first adopt the supervised contrastive training strategy during the pre-training phase to fully explore potential contrast patterns of weakly-labeled data to learn robust representations. Then we design a simple dropping-layer strategy to remove the top layers from the pre-trained model that are susceptible to noisy data. Last, we add a classification layer on top of the remaining model and fine tune it with labeled data. The proposed framework can learn rich contrastive sentiment patterns in the case of label noise and is applicable to a variety of deep encoders. The experimental results on the Amazon product review, Twitter and SST5 datasets demonstrate the superiority of our method.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8

Similar content being viewed by others

Data availibility statement

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Notes

  1. One category has a fixed possibility of being labeled as another.

  2. We use the pre-trained bert-base with 12 layers and 768 hidden units, https://storage.googleapis.com/bert_models/2020_02_20/uncased_L-12_H-768_A-12.zip.

  3. Emojis are not used as an input to train the network.

  4. http://help.sentiment140.com/for-students/.

  5. https://github.com/datastories-semeval2017-task4.

  6. http://nlp.stanford.edu/sentiment.

  7. The scores of the two different metrics are slightly different because of the equal number of samples in each class of all test sets. The macro-F1 is taking the average of the F1 scores calculated from each class, thus regarding each category equally.

References

  1. Turney PD (2002) Thumbs up or thumbs down? Semantic orientation applied to unsupervised classification of reviews. In: Proceedings of the 40th annual meeting of the association for computational linguistics, pp 421–439

  2. Hu M, Liu B (2004) Mining and summarizing customer reviews. In: Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp 168–177. https://doi.org/10.1145/1014052.1014073

  3. Pang B, Lee L, Vaithyanathan S (2002) Thumbs up? Sentiment classification using machine learning techniques. In: Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing, pp 79–86

  4. Li C, Gao F, Bu J, Xu L, Chen X, Gu Y, Shao Z, Zheng Q, Zhang N, Wang Y, Yu Z (2021) SentiPrompt: sentiment knowledge enhanced prompt-tuning for aspect-based sentiment analysis. CoRR arXiv:2109.08306

  5. Timo S, Hinrich S (2021) Exploiting cloze-questions for few-shot text classification and natural language inference. In: Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics, pp 255–269

  6. Zhao W, Guan Z, Chen L, He X, Deng D, Wang B, Wang Q (2017) Weakly-supervised deep embedding for product review sentiment analysis. IEEE Trans Knowl Data Eng 30(1):185–197. https://doi.org/10.1109/TKDE.2017.2756658

    Article  Google Scholar 

  7. Eric X, Michael J, Stuart JR, Andrew N (2002) Distance metric learning with application to clustering with side-information. Adv Neural Inf Process Syst 15:521–528

    Google Scholar 

  8. Kristina T, Anna R, Luke Z, Dilek H, Iz B, Steven B, Ryan C, Tanmoy C, Zhou Y (2021) Few-shot text classification with triplet networks, data augmentation,and curriculum learning. In: Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics, pp 5493–5500

  9. Guan Z, Chen L, Zhao W, Zheng Y, Tan S, Deng C (2016) Weakly-supervised deep learning for customer review sentiment classification. In: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, pp 3719–3725

  10. Ting C, Simon K, Mohammad N, Geoffrey H (2020) A simple framework for contrastive learning of visual representations. In: Proceedings of the 37th International Conference on Machine Learning, pp 1597–1607

  11. John MG, Osvald N, Gary DB, Bo W (2021) Declutr: Deep contrastive learning for unsupervised textual representations. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pp 879–895. https://doi.org/10.18653/v1/2021.acl-long.72

  12. Aritra G, Andrew L (2021) Contrastive learning improves model robustness under label noise. In: IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp 2703–2708

  13. Xiao T, Xia T, Yang Y, Chang H, Wang X (2015) Learning from massive noisy labeled data for image classification. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2691–2699. https://doi.org/10.1109/CVPR.2015.7298885

  14. Goldberger J, Ben-Reuven E (2017) Training deep neural-networks using a noise adaptation layer. In: Proceedings of the 5th International Conference on Learning Representations

  15. Ishan J, Matthew N, Xuewen C (2016) Learning deep networks from noisy labels with dropout regularization. In: 2016 IEEE 16th International Conference on Data Mining, pp 67–972. https://doi.org/10.1109/ICDM.2016.0121

  16. Ghosh A, Kumar H, Sastry PS (2017) Robust loss functions under label noise for deep neural networks. In: Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence, pp 1919–1925

  17. Zhang Z, Sabuncu MR (2018) Generalized cross entropy loss for training deep neural networks with noisy labels. In: Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems 2018, pp 8792–8802

  18. Zhang C, Samy B, Moritz H, Benjamin R, Oriol V (2021) Understanding deep learning (still) requires rethinking generalization. Commun ACM 64(3):107–115. https://doi.org/10.1145/3446776

    Article  Google Scholar 

  19. Li J, Zhang M, Xu K, Dickerson PJ, Ba J (2020) Noisy labels can induce good representations. CoRR arXiv:abs/2012.12896

  20. Liu H, Dai Z, David R, Quoc V (2021) Pay attention to MLPs. CoRR arXiv:abs/2105.08050

  21. Zhang S, Xu X, Pang Y, Han J (2020) Multi-layer attention based CNN for target-dependent sentiment classification. Neural Process Lett 51(3):2089–2103. https://doi.org/10.1007/s11063-019-10017-9

    Article  Google Scholar 

  22. Habimana O, Li Y, Li R, Gu X, Yan W (2020) Attentive convolutional gated recurrent network: a contextual model to sentiment analysis. Int J Mach Learn Cyber 11:2637–2651. https://doi.org/10.1007/s13042-020-01135-1

    Article  Google Scholar 

  23. Al-Smadi M, Talafha B, Al-Ayyoub M, Jararweh Y (2019) Using long short-term memory deep neural networks for aspect-based sentiment analysis of Arabic reviews. Int J Mach Learn Cyber 10:2163–2175. https://doi.org/10.1007/s13042-018-0799-4

    Article  Google Scholar 

  24. Arunava KC, Sourav D, Anup KK (2021) Sentiment analysis of Covid-19 tweets using evolutionary classification-based LSTM model. CoRR arXiv:abs/2106.06910

  25. Ling M, Chen Q, Sun Q, Jia Y (2020) Hybrid neural network for Sina Weibo sentiment analysis. IEEE Trans Comput Soc Syst 7(4):983–990

    Article  Google Scholar 

  26. Devlin J, Chang M, Lee K, Toutanova K (2019) BERT: pre-training of deep bidirectional transformers for language understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, pp 4171–4186. https://doi.org/10.18653/v1/n19-1423

  27. Ashish V, Noam S, Niki P, Jakob U, Llion J, Aidan NG, Łukasz K, Illia P (2017) Attention is all you need. In: Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, pp 998–6008

  28. Alec G, Richa B, Huang L (2009) Twitter sentiment classification using distant supervision. CS224N Project Report Stanford. https://doi.org/10.1109/COMSNETS.2017.7945451

    Article  Google Scholar 

  29. Qu L, Gemulla R, Weikum G (2012) A weakly supervised model for sentence-level semantic orientation analysis with multiple experts. In: Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, pp 149–159

  30. Täckström O, McDonald RT (2011) Semi-supervised latent variable models for sentence-level sentiment analysis. In: The 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference, pp 569–574

  31. Wang B, Shan D, Fan A, Liu L, Gao J (2022) A sentiment classification method of web social media based on multidimensional and multilevel modeling. IEEE Trans Ind Informatics 18(2):1240–1249

    Article  Google Scholar 

  32. Wang F, Liu H (2021) Understanding the behaviour of contrastive loss. In: IEEE Conference on Computer Vision and Pattern Recognition, pp 2495–2504

  33. Chen X, Gupta A (2015) Webly supervised learning of convolutional networks. In: 2015 IEEE International Conference on Computer Vision, pp 1431–1439. https://doi.org/10.1109/ICCV.2015.168

  34. Alec G, Richa B, Huang (2014) Training convolutional networks with noisy labels. CoRR abs/1406.2080

  35. Nitish S, Geoffrey H, Alex K, Ilya S, Ruslan S (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15(1):1929–1958

    MathSciNet  MATH  Google Scholar 

  36. Bekker AJ, Goldberger J (2016) Training deep neural-networks based on unreliable labels. In: 2016 IEEE International Conference on Acoustics, Speech and Signal Processing, pp 2682–2686. https://doi.org/10.1109/ICASSP.2016.7472164

  37. Cheng L, Zhou X, Zhao L, Li D, Shang H, Zheng Y, Pan P, Xu Y (2020) Weakly supervised learning with side information for noisy labeled images. In: European Conference on Computer Vision, pp 306–321. https://doi.org/10.1007/978-3-030-58577-8_19

  38. Naresh M, PS S (2013) Noise tolerance under risk minimization. IEEE Trans Cybern 43(3):1146–1151. https://doi.org/10.1109/TSMCB.2012.2223460

    Article  Google Scholar 

  39. Khosla P, Teterwak P, Wang C, Sarna A, Tian Y, Isola P, Maschinot A, Liu C, Krishnan D (2020) Supervised contrastive learning. In: Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020

  40. Yosinski J, Clune J, Bengio Y, Lipson H (2014) How transferable are features in deep neural networks?. In: Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Processing Systems 2014, pp 3320–3328

  41. Socher R, Perelygin A, Wu J, Chuang J, Manning DC, Andrew Y, Potts C (2013) Recursive deep models for semantic compositionality over a sentiment treebank. In: Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pp 1631–1642

  42. Ding X, Liu B, Philip SY (2008) A holistic lexicon-based approach to opinion mining. In: Proceedings of the 2008 international conference on web search and data mining, pp 231–240. https://doi.org/10.1145/1341531.1341561

  43. Wang S, Christopher DM (2012) Baselines and bigrams: simple, good sentiment and topic classification. In: Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics, pp 90–94

  44. Tang D, Wei F, Nan Y, Ming Z, Bing Q (2014) Learning sentiment-specific word embedding for twitter sentiment classification. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, pp 1555–1565. https://doi.org/10.3115/v1/p14-1146

  45. Gunel B, Du J, Conneau A, Stoyanov V (2021) Supervised contrastive learning for pre-trained language model fine-tuning. In: 9th International Conference on Learning Representations

  46. Ilya T, Neil H, Alexander K, Lucas B, Zhai X, Thomas U, Jessica Y, Daniel K, Jakob U, Mario L, Dosovitskiy A (2021) Mlp-mixer: an all-mlp architecture for vision. CoRR arXiv:abs/2105.01601

  47. Tomas M, Ilya S, Chen K, Greg S, Jeff D (2013) Distributed representations of words and phrases and their compositionality. In: Advances in Neural Information Processing Systems 26: 27th Annual Conference on Neural Information Processing Systems 2013, pp 3111–3119

  48. Hu Z, Wu H, Liao S, Hu H, Liu S, Li B (2018) Person re-identification with hybrid loss and hard triplets mining. In: Fourth IEEE International Conference on Multimedia Big Data, pp 1–5. https://doi.org/10.1109/BigMM.2018.8499463

  49. Maaten L, Hinton G (2008) Visualizing data using t-SNE. J Mach Learn Res 9(2605):2579–2605

    MATH  Google Scholar 

Download references

Acknowledgements

This research was supported by the National Natural Science Foundation of China (Grant Nos. 61902316, 62133012, 61936006, 61876144, 61876145, 62073255, 62103314, 61973249, 62001381), the Key Research and Development Program of Shaanxi (Program Nos. 2020ZDLGY04-07, 2021ZDLGY02-06), Innovation Capability Support Program of Shaanxi (Program No. 2021TD-05), and Natural Science Basic Research Program of Shaanxi (Program Nos. 2022JQ-675, 2021JQ-712).

Author information

Author notes

  1. Long Chen and Fei Wang contributed equally to this work.

Authors and Affiliations

  1. Shaanxi Key Laboratory of Information Communication Network and Security, Xi’an University of Posts and Telecommunications, Xi’an, China

    Long Chen, Fei Wang & Wenjing Wang

  2. School of Information Science and Technology, Northwest University, Xi’an, China

    Ruijing Yang

  3. Academy of Advanced Interdisciplinary Research, Xidian University, Xi’an, China

    Fei Xie

  4. School of Computer Science and Technology, Xidian University, Xi’an, China

    Cai Xu, Wei Zhao & Ziyu Guan

Authors
  1. Long Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Fei Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Ruijing Yang
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Fei Xie
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Wenjing Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Cai Xu
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Wei Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Ziyu Guan
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Fei Xie.

Ethics declarations

Conflict of interest

The authors declare that they have no confict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, L., Wang, F., Yang, R. et al. Representation learning from noisy user-tagged data for sentiment classification. Int. J. Mach. Learn. & Cyber. 13, 3727–3742 (2022). https://doi.org/10.1007/s13042-022-01622-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s13042-022-01622-7

Keywords

Search

Navigation