Abstract
Detecting and aggregating sentiments toward people, organizations, and events expressed in unstructured social media have become critical text mining operations. Early systems detected sentiments over whole passages, whereas more recently, target-specific sentiments have been of greater interest. In this paper, we present MTTDSC, a multi-task target-dependent sentiment classification system that is informed by feature representation learnt for the related auxiliary task of passage-level sentiment classification. The auxiliary task uses a gated recurrent unit (GRU) and pools GRU states, followed by an auxiliary fully-connected layer that outputs passage-level predictions. In the main task, these GRUs contribute auxiliary per-token representations over and above word embeddings. The main task has its own, separate GRUs. The auxiliary and main GRUs send their states to a different fully connected layer, trained for the main task. Extensive experiments using two auxiliary datasets and three benchmark datasets (of which one is new, introduced by us) for the main task demonstrate that MTTDSC outperforms state-of-the-art baselines. Using word-level sensitivity analysis, we present anecdotal evidence that prior systems can make incorrect target-specific predictions because they miss sentiments expressed by words independent of target.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
MTTDSC code and datasets are available at https://github.com/divamgupta/mttdsc.
- 2.
We elide possible scalar offsets in sigmoids and softmaxes for simplicity throughout the paper.
- 3.
References
Baccianella, S., Esuli, A., Sebastiani, F.: Evaluation measures for ordinal regression. In: Intelligent Systems Design and Applications, Pisa, Italy, pp. 283–287 (2009)
Bengio, Y., Louradour, J., Collobert, R., Weston, J.: Curriculum learning. In: ICML, Montreal, Canada, pp. 41–48 (2009)
Choi, E., Hewlett, D., Uszkoreit, J., Polosukhin, I., Lacoste, A., Berant, J.: Coarse-to-fine question answering for long documents. In: ACL, pp. 209–220 (2017)
Chung, J., Gulcehre, C., Cho, K., Bengio, Y.: Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555 (2014)
Dong, L., Wei, F., Tan, C., Tang, D., Zhou, M., Xu, K.: Adaptive recursive neural network for target-dependent Twitter sentiment classification. In: ACL, Baltimore, Maryland, USA, pp. 49–54 (2014)
Esuli, A., Sebastiani, F.: SentiWordNet: a publicly available lexical resource for opinion mining. In: LREC, pp. 417–422 (2006)
Go, A., Bhayani, R., Huang, L.: Twitter sentiment classification using distant supervision. CS224N Project Report, Stanford, vol. 1, no. 2009, p. 12 (2009)
Godbole, N., Srinivasaiah, M., Skiena, S.: Large-scale sentiment analysis for news and blogs. In: ICWSM, pp. 219–222 (2007)
Jiang, L., Yu, M., Zhou, M., Liu, X., Zhao, T.: Target-dependent twitter sentiment classification. In: ACL, Portland, Oregon, USA, pp. 151–160 (2011)
Joachims, T.: Optimizing search engines using clickthrough data. In: SIGKDD Conference, pp. 133–142. ACM (2002)
Kim, Y.: Convolutional neural networks for sentence classification. arXiv preprint arXiv:1408.5882 (2014)
Liu, B.: Sentiment analysis and opinion mining. Synth. Lect. Hum. Lang. Technol. 5(1), 1–167 (2012)
Maurer, A., Pontil, M., Romera-Paredes, B.: The benefit of multitask representation learning. J. Mach. Learn. Res. 17(81), 1–32 (2016)
Pak, A., Paroubek, P.: Twitter as a corpus for sentiment analysis and opinion mining. In: LREC, Valletta, Malta, pp. 1–3 (2010)
Pang, B., Lee, L., et al.: Opinion mining and sentiment analysis. Found. Trends Inf. Retrieval 2(1–2), 1–135 (2008)
Peng, H., Thomson, S., Smith, N.A.: Deep multitask learning for semantic dependency parsing. arXiv preprint arXiv:1704.06855 (2017)
Pennington, J., Socher, R., Manning, C.D.: GloVe: global vectors for word representation. In: EMNLP Conference 2014, pp. 1532–1543 (2014)
Ruder, S., Bingel, J., Augenstein, I., Søgaard, A.: Learning what to share between loosely related tasks. arXiv preprint arXiv:1705.08142 (2017)
Sanders, N.: Twitter sentiment corpus (2011)
Tang, D., Qin, B., Feng, X., Liu, T.: Effective LSTMs for target-dependent sentiment classification. arXiv preprint arXiv:1512.01100 (2015)
Teng, Z., Vo, D.-T., Zhang, Y.: Context-sensitive lexicon features for neural sentiment analysis. In: EMNLP, Austin, Texas, USA, pp. 1629–1638 (2016)
Wang, B., Liakata, M., Zubiaga, A., Procter, R.: TDParse: multi-target-specific sentiment recognition on Twitter. In: EACL, pp. 483–493 (2017)
Wilson, T., Wiebe, J., Hoffmann, P.: Recognizing contextual polarity in phrase-level sentiment analysis. In: EMNLP, Vancouver, BC, Canada, pp. 347–354 (2005)
Acknowledgement
The project was partially supported by IBM, Early Career Research Award (SERB, India), and the Center for AI, IIIT Delhi, India.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Gupta, D., Singh, K., Chakrabarti, S., Chakraborty, T. (2019). Multi-task Learning for Target-Dependent Sentiment Classification. In: Yang, Q., Zhou, ZH., Gong, Z., Zhang, ML., Huang, SJ. (eds) Advances in Knowledge Discovery and Data Mining. PAKDD 2019. Lecture Notes in Computer Science(), vol 11439. Springer, Cham. https://doi.org/10.1007/978-3-030-16148-4_15
Download citation
DOI: https://doi.org/10.1007/978-3-030-16148-4_15
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-16147-7
Online ISBN: 978-3-030-16148-4
eBook Packages: Computer ScienceComputer Science (R0)