Semi-supervised SRL System with Bayesian Inference

  • Alejandra Lorenzo
  • Christophe Cerisara
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8403)

Abstract

We propose a new approach to perform semi-supervised training of Semantic Role Labeling models with very few amount of initial labeled data. The proposed approach combines in a novel way supervised and unsupervised training, by forcing the supervised classifier to overgenerate potential semantic candidates, and then letting unsupervised inference choose the best ones. Hence, the supervised classifier can be trained on a very small corpus and with coarse-grain features, because its precision does not need to be high: its role is mainly to constrain Bayesian inference to explore only a limited part of the full search space. This approach is evaluated on French and English. In both cases, it achieves very good performance and outperforms a strong supervised baseline when only a small number of annotated sentences is available and even without using any previously trained syntactic parser.

Keywords

Entropy Boulder Core Role 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Màrquez, L., Carreras, X., Litkowski, K.C., Stevenson, S.: Semantic role labeling: An introduction to the special issue. Comput. Linguist. 34, 145–159 (2008)CrossRefGoogle Scholar
  2. 2.
    Pradhan, S.S., Ward, W., Martin, J.H.: Towards robust semantic role labeling. Comput. Linguist. 34, 289–310 (2008)CrossRefGoogle Scholar
  3. 3.
    He, S., Gildea, H.: Self-training and Cotraining for Semantic Role Labeling: Primary Report. Technical report, TR 891, University of Colorado at Boulder (2006)Google Scholar
  4. 4.
    Lee, J.Y., Song, Y.I., Rim, H.C.: Investigation of weakly supervised learning for semantic role labeling. In: ALPIT, pp. 165–170 (2007)Google Scholar
  5. 5.
    Daumé III, H.: Semi-supervised or semi-unsupervised? In: Proc. NAACL Wokshop on Semi-supervised Learning for NLP (2009)Google Scholar
  6. 6.
    Titov, I., Klementiev, A.: Semi-supervised semantic role labeling: Approaching from an unsupervised perspective. In: Proceedings of the International Conference on Computational Linguistics (COLING), Bombay, India (2012)Google Scholar
  7. 7.
    Jain, D., Beetz, M.: Soft evidential update via markov chain monte carlo inference. In: Dillmann, R., Beyerer, J., Hanebeck, U.D., Schultz, T. (eds.) KI 2010. LNCS, vol. 6359, pp. 280–290. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  8. 8.
    Palmer, M., Gildea, D., Kingsbury, P.: The proposition bank: An annotated corpus of semantic roles. Comput. Linguist. 31, 71–106 (2005)CrossRefGoogle Scholar
  9. 9.
    Dowty, D.: Thematic proto-roles and argument selection. Language 67, 547–619 (1991)CrossRefGoogle Scholar
  10. 10.
    Bohnet, B.: Top accuracy and fast dependency parsing is not a contradiction. In: Proc. International Conference on Computational Linguistics, Beijing, China (2010)Google Scholar
  11. 11.
    Björkelund, A., Hafdell, L., Nugues, P.: In: Proceedings of the Thirteenth Conference on Computational Natural Language Learning: Shared Task, CoNLL 2009, pp. 43–48. Association for Computational Linguistics, Stroudsburg (2009)CrossRefGoogle Scholar
  12. 12.
    Deschacht, K., Moens, M.F.: Semi-supervised semantic role labeling using the latent words language model. In: EMNLP, pp. 21–29 (2009)Google Scholar
  13. 13.
    van der Plas, L., Merlo, P., Henderson, J.: Scaling up automatic cross-lingual semantic role annotation. In: Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies: Short Papers, HLT 2011, vol. 2, pp. 299–304. Association for Computational Linguistics (2011)Google Scholar
  14. 14.
    Surdeanu, M., Johansson, R., Meyers, A., Màrquez, L., Nivre, J.: The conll-2008 shared task on joint parsing of syntactic and semantic dependencies. In: Proceedings of the Twelfth Conference on Computational Natural Language Learning, CoNLL 2008, pp. 159–177. Association for Computational Linguistics, Stroudsburg (2008)CrossRefGoogle Scholar
  15. 15.
    Zhu, X.: Semi-Supervised Learning Literature Survey. Technical report, Computer Sciences, University of Wisconsin-Madison (2005)Google Scholar
  16. 16.
    Pise, N.N., Kulkarni, P.: A survey of semi-supervised learning methods. In: Proceedings of the 2008 International Conference on Computational Intelligence and Security, CIS 2008, vol. 2, pp. 30–34. IEEE Computer Society, Washington, DC (2008)CrossRefGoogle Scholar
  17. 17.
    Fürstenau, H., Lapata, M.: Graph alignment for semi-supervised semantic role labeling. In: EMNLP, pp. 11–20 (2009)Google Scholar
  18. 18.
    Das, D., Smith, N.A.: Semi-supervised frame-semantic parsing for unknown predicates. In: Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, HLT 2011, vol. 1, pp. 1435–1444. Association for Computational Linguistics, Stroudsburg (2011)Google Scholar
  19. 19.
    Haghighi, A., Klein, D.: Prototype-driven learning for sequence models. In: Proceedings of the Main Conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics, HLT-NAACL 2006, pp. 320–327. Association for Computational Linguistics, Stroudsburg (2006)CrossRefGoogle Scholar
  20. 20.
    Haghighi, A., Klein, D.: Prototype-driven grammar induction. In: Proceedings of the 21st International Conference on Computational Linguistics and the 44th Annual Meeting of the Association for Computational Linguistics. ACL-44, pp. 881–888. Association for Computational Linguistics, Stroudsburg (2006)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Alejandra Lorenzo
    • 1
  • Christophe Cerisara
    • 1
  1. 1.LORIA / UMR 7503Vandoeuvre-les-NancyFrance

Personalised recommendations