Advertisement

Convolution Neural Network with Active Learning for Information Extraction of Enterprise Announcements

  • Lei Fu
  • Zhaoxia Yin
  • Yi Liu
  • Jun ZhangEmail author
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11109)

Abstract

We propose using convolution neural network (CNN) with active learning for information extraction of enterprise announcements. The training process of supervised deep learning model usually requires a large amount of training data with high-quality reference samples. Human production of such samples is tedious, and since inter-labeler agreement is low, very unreliable. Active learning helps assuage this problem by automatically selecting a small amount of unlabeled samples for humans to hand correct. Active learning chooses a selective set of samples to be labeled. Then the CNN is trained on the labeled data iteratively, until the expected experimental effect is achieved. We propose three sample selection methods based on certainty criterion. We also establish an enterprise announcements dataset for experiments, which contains 10410 samples totally. Our experiment results show that the amount of labeled data needed for a given extraction accuracy can be reduced by more than 45.79% compared to that without active learning.

Keywords

Text classification Active learning Convolutional neural networks Enterprise announcements 

Notes

Acknowledgments

This work is partially supported by Shenzhen Science & Research projects (No: JCYJ20160331104524983) and Key Technologies Research & Development Program of Shenzhen (No: JSGG20160229121006579). We thank the reviewers for the insightful comments.

References

  1. 1.
    Yogatama, D., et al.: Generative and discriminative text classification with recurrent neural networks (2017). arXiv preprint, arXiv:1703.01898
  2. 2.
    Tu, Z., et al.: Modeling coverage for neural machine translation (2016). arXiv preprint, arXiv:1601.04811
  3. 3.
    Cao, Z., et al.: Improving multi-document summarization via text classification. In: AAAI (2017)Google Scholar
  4. 4.
    Conneau, A., et al.: Very deep convolutional networks for natural language processing (2016). arXiv preprint, arXiv:1606.01781
  5. 5.
    Joshi, A.J. Porikli, A.J., Papanikolopoulos, N.: Multi-class active learning for image classification. In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2009, pp. 2372–2379 (2009)Google Scholar
  6. 6.
    Tur, G., Hakkani, D.: Combining active and semi-supervised learning for spoken language understanding. Speech Commun. 45(2), 171–186 (2005)CrossRefGoogle Scholar
  7. 7.
    Settles, B.: Active learning literature survey. Univ. Wisconsin, Madison 52(55–66), 11 (2010)Google Scholar
  8. 8.
    MacKay, David J.C.: Information-based objective functions for active data selection. Neural Comput. 4(4), 590–604 (1992)CrossRefGoogle Scholar
  9. 9.
    Kapoor, A., Grauman, K., Urtasun, R., Darrell, T.: Active learning with Gaussian processes for object categorization. In: ICCV (2007)Google Scholar
  10. 10.
    Roy, N., McCallum, A.: Toward optimal active learning through monte carlo estimation of error reduction. In: ICML (2001)Google Scholar
  11. 11.
    Gal, Y., Ghahramani, Z.: Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In: International Conference on Machine Learning (2016)Google Scholar
  12. 12.
    Li, X., Guo, Y.: Active learning with multi-label SVM classification. In: IJCAI (2013)Google Scholar
  13. 13.
    Seung, H.S., Opper, M., Sompolinsky, H.: Query by committee. In: Proceedings of the Fifth Annual Workshop on Computational Learning Theory. ACM (1992)Google Scholar
  14. 14.
    McCallumzy, A.K., Nigamy, K.: Employing EM and pool-based active learning for text classification. In: Proceedings of the International Conference on Machine Learning (ICML), pp.359–367 (1998). CiteseerGoogle Scholar
  15. 15.
    Wang, K., et al.: Cost-effective active learning for deep image classification. IEEE Trans. Circ. Syst. Video Technol. 27(12), 2591–2600 (2017)CrossRefGoogle Scholar
  16. 16.
    Dasgupta, S.: Coarse sample complexity bounds for active learning. In: Advances in Neural Information Processing Systems (2006)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Key Laboratory of Intelligent Computing and Signal Processing, Ministry of EducationAnhui UniversityHefeiPeople’s Republic of China
  2. 2.PKU Shenzhen InstituteShenzhenChina
  3. 3.Shenzhen Securities Information, Co., Ltd.ShenzhenChina

Personalised recommendations