Elamur Research⌕
☰
home / papers / 2010 / Kernel Based Discourse Relation Recognition with Temporal Ordering Information

Kernel Based Discourse Relation Recognition with Temporal Ordering Information

Summary

The paper addresses automatic discourse relation recognition: finding which text spans are linked and assigning the relation type. It covers explicit relations, which are signalled by a connective such as "but", and implicit ones. Earlier systems encoded syntax only through hand-chosen flat paths or 2-level production rules, which discard much of the parse-tree structure. The authors instead apply a convolution tree kernel directly to parse trees inside an SVM. They merge it in a composite kernel with conventional flat features such as lexical pairs, argument distance and attribution. They also add tense and temporal ordering features to narrow down plausible relation readings, drawing on Webber's (1988) view of tense as discourse anaphor. Experiments on PDTB 2.0 use the four top-level sense classes. The tree kernel gives statistically significant gains over the flat syntactic path feature, and the temporal information adds further significant improvements for both relation types. The work matters because it reduces manual syntactic feature engineering for discourse analysis.

Key points

  • Data: Penn Discourse Treebank (PDTB) 2.0, covering 2,312 Wall Street Journal articles. Only the top-level senses are used: Temporal, Contingency, Comparison and Expansion. AltLex cases are handled like implicit relations.
  • Unlike 2-level production rules (Lin et al., 2009), the tree kernel also captures deeper sub-tree structures. This places the model in a higher-dimensional feature space.
  • Framework: separate binary SVM classifiers are trained per relation type, with negative examples drawn from neighbouring non-argument clauses. At test time, candidate pairs come from within a sentence or a three-sentence window, and each pair gets the most confident label.
  • The combined model (base features + tree kernel + tense/temporal features) reaches an accuracy of 74.2% on explicit relations, 40.0% on implicit relations and 57.3% overall.
  • Planned future work is modelling Entity relations, which account for 24% of the Implicit+Entity+No relation cases.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Discourse analysis

Cited by

16 papers of our library cite it (of the 140673 whose bibliographies we read); 9 of them have a page here, 7 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{wang2010kernel,
  title = {{Kernel Based Discourse Relation Recognition with Temporal Ordering Information}},
  author = {Wang, WenTing and Su, Jian and Tan, Chew Lim},
  booktitle = {Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics},
  year = {2010},
  url = {https://aclanthology.org/P10-1073/}
}

References

19 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ben W. and James P. 2007. Automatically Identifying the Arguments of Discourse Connectives. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, pages 92-101. find
  2. Culotta A. and Sorensen J. 2004. Dependency Tree Kernel for Relation Extraction. In Proceedings of the 40 th Annual Meeting of the Association for Computational Linguistics (ACL 2004), pages 423-429. find
  3. Collins M. and Duffy N. 2001. New Ranking Algorithms for Parsing and Tagging: Kernels over Discrete Structures and the Voted Perceptron. In Proceedings of the 40 th Annual Meeting of the Association for Computational Linguistics (ACL 2002), pages 263-270. find
  4. Collins M. and Duffy N. 2002. Convolution Kernels for Natural Language. NIPS-2001. find
  5. Haussler D. 1999. Convolution Kernels on Discrete Structures. Technical Report UCS-CRL-99-10, University of California, Santa Cruz. find
  6. Joachims T. 1999. Making Large-scale SVM Learning Practical. In Advances in Kernel Methods – Support Vector Learning. MIT Press. find
  7. Knott, A., Oberlander, J., O’Donnel, M., and Mellish, C. 2001. Beyond elaboration: the interaction of relations and focus in coherent text. In T. Sanders, J. Schilperoord, and W. Spooren, editors, Text Representation: Linguistic and Psycholinguistics Aspects, pages 181-196. Benjamins, Amsterdam. find
  8. Lee A., Prasad R., Joshi A., Dinesh N. and Webber B. 2006. Complexity of dependencies in discourse: are dependencies in discourse more complex than in syntax? In Proceedings of the 5th International Workshop on Treebanks and Linguistic Theories. Prague, Czech Republic, December. find
  9. Lin Z., Kan M. and Ng H. 2009. Recognizing Implicit Discourse Relations in the Penn Discourse Treebank. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing (EMNLP 2009), Singapore, August. find our card
  10. Marcu D. and Echihabi A. 2002. An Unsupervised Approach to Recognizing Discourse Relations. In Proceedings of the 40 th Annual Meeting of ACL, pages 368-375. find our card
  11. Moschitti A. 2004. A Study on Convolution Kernels for Shallow Semantic Parsing. In Proceedings of the 42 th Annual Meeting of the Association for Computational Linguistics (ACL 2004), pages 335-342. find
  12. Pettibone J. and Pon-Barry H. 2003. A Maximum Entropy Approach to Recognizing Discourse Relations in Spoken Language. Working Paper. The Stanford Natural Language Processing Group, June 6. find
  13. Pitler E., Louis A. and Nenkova A. 2009. Automatic Sense Predication for Implicit Discourse Relations in Text. In Proceedings of the Joint Conference of the 47 th Annual Meeting of the Association for Computational Linguistics and the 4 th International Joint Conference on Natural Language Processin… find
  14. Prasad R., Dinesh N., Lee A., Miltsakaki E., Robaldo L., Joshi A. and Webber B. 2008. The Penn Discourse TreeBank 2.0. In Proceedings of the 6 th International Conference on Language Resources and Evaluation (LREC 2008). find our card
  15. Saito M., Yamamoto K. and Sekine S. 2006. Using phrasal patterns to identify discourse relations. In Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics (HLT-NAACL 2006), pages 133–136, New York, USA. find our card
  16. Vapnik V. 1995. The Nature of Statistical Learning Theory. Springer-Verlag, New York. find
  17. Webber Bonnie. 1988. Tense as Discourse Anaphor. Computational Linguistics, 14:61–73. find
  18. Zelenko D., Aone C. and Richardella A. 2003. Kernel Methods for Relation Extraction. Journal of Machine Learning Research, 3(6):1083-1106. find
  19. Zhang M., Zhang J. and Su J. Exploring Syntactic Features for Relation Extraction using a Convolution Tree Kernel. In Proceedings of the Human Language Technology conference - North American chapter of the Association for Computational Linguistics annual meeting (HLT-NAACL 2006), New York, USA. find

↑ back to top

This page in Markdown →