Elamur Research⌕
☰
home / papers / 2012 / Global Features for Shallow Discourse Parsing

Global Features for Shallow Discourse Parsing

Summary

The paper addresses shallow discourse parsing in the PDTB 2.0 framework. Given an explicit connective, the system must find the exact spans of its two arguments, Arg1 and Arg2. An earlier CRF-based parser (Ghosh et al., 2011a) relied on token-level features and a narrow context window. As a result, it often missed Arg1, which is not syntactically tied to the connective and may sit in a different sentence.

The authors keep that parser as a first stage and use it to produce n-best candidate analyses over a five-sentence window around each connective. A second stage reranks these candidates with a compact set of hand-crafted, data-driven global features that need no syntactic parse tree. Examples include the CRF log posterior, signals for missing or surplus arguments, Arg1 position relative to the connective sentence, and checks on boundary tags.

On PDTB with gold connectives, reranking significantly beats the previous best parser. The gain comes mainly from better Arg1 recall at a modest precision cost, while Arg2 hardly changes. This shows that discourse-level structure can be added cheaply through reranking.

Key points

  • Data: Penn Discourse TreeBank 2.0, about one million words of Wall Street Journal text with 18,459 explicit connective tokens (100 types). Only the four top-level senses are used (TEMPORAL, CONTINGENCY, COMPARISON, EXPANSION).
  • Arg1 is the harder target: it occurs in the connective's sentence 60.9% of the time, in the adjacent previous sentence 30.1%, and in a non-adjacent earlier sentence 9.0%.
  • In greedy feature selection for the Arg1 reranker (k = 16, development set, exact match), F1 rises from 59.64 with the posterior alone (GF0) to 62.30 with GF0+GF2+GF1+GF4. The undergeneration feature (GF2) is the only one that helps on its own.
  • For inter-sentential Arg1 with an online passive-aggressive perceptron reranker, exact-match F1 goes from 36.44 to 37.56 and overlap F1 from 54.88 to 56.23, with recall improving while precision falls.
  • Reranker gains level off by 16-best lists even though oracle scores keep rising. The authors suggest the parse-independent features could also apply to dialogue data annotated in PDTB style.

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

10 papers of our library cite it (of the 140673 whose bibliographies we read); 3 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{ghosh2012global,
  title = {{Global Features for Shallow Discourse Parsing}},
  author = {Ghosh, Sucheta and Riccardi, Giuseppe and Johansson, Richard},
  booktitle = {Proceedings of the 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
  year = {2012},
  url = {https://aclanthology.org/W12-1622/}
}

References

32 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Regina Barzilay, Lillian Lee, et al. 2004. Catching the drift: Probabilistic content models, with applications to generation and summarization. In Proc. of NAACL-HLT, 2004. find
  2. Regina Barzilay, Mirella Lapata, et al. 2005. Modeling local coherence: an entity-based approach. In Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics (ACL05). find
  3. E. Charniak and M. Johnson. 2005. Coarse-to-fine n-best parsing and maxent discriminative reranking. In Proceedings of the 43rd Annual Meeting of the ACL. find our card
  4. Michael Collins and Nigel Duffy. 2002. New ranking algorithms for parsing and tagging: Kernels over discrete structures, and the voted perceptron. In ACL02. find
  5. Michael Collins. 2000. Discriminative reranking for natural language parsing. In Computational Linguistics, pages 175–182. Morgan Kaufmann. find our card
  6. Michael Collins. 2002. Ranking algorithms for named-entity extraction: Boosting and the voted perceptron. In Proceedings of ACL 2002. find
  7. Koby Crammer, Ofer Dekel, Joseph Keshet, Shai Shalev-Schwartz, and Yoram Singer. 2006. Online passive-aggressive algorithms. Journal of Machine Learning Research, 7:551–585. find
  8. Hal Daume III, Daniel Marcu, et al. 2004. Np bracketing by maximum entropy tagging and svm reranking. In Proceedings of EMNLP’04. find
  9. Rong-En Fan, Chih-Jen Lin, Kai-Wei Chang, Xiang-Rui Wang, et al. 2008. Liblinear: A library for large linear classification. Journal of Machine Learning Research. find
  10. Sucheta Ghosh, Richard Johansson, Giuseppe Riccardi, and Sara Tonelli. 2011a. Shallow discourse parsing with conditional random fields. In Proceedings of the 5th International Joint Conference on Natural Language Processing (IJC-NLP), Chiang Mai, Thailand. find our card
  11. Sucheta Ghosh, Sara Tonelli, Giuseppe Riccardi, and Richard Johansson. 2011b. End-to-end discourse parser evaluation. In Proceedings of 5th IEEE International Conference on Semantic Computing, Palo Alto, CA, USA. find
  12. Sucheta Ghosh, Richard Johansson, Giuseppe Riccardi, and Sara Tonelli. 2012. Improving the recall of a discourse parser by constraint-based postprocessing. In Proceedings of International Conference on Languages Resources and Evaluations (LREC 2012). find
  13. B.J. Grosz, J. Hircshberg, et al. 1992. Some intonational characteristics of discourse structure. In Ohala et al., editors, Proceedings of the International Conference on Spoken Language Processing, Vol. 1, volume 1, pages 429–432. find
  14. B.J. Grosz, A. K. Joshi, S. Weinstein, et al. 1995. Centering: A framework for modeling the local coherence of discourse. Computational Linguistics, 21(2). find our card
  15. J. S. Urban Hjorth. 1993. Computer Intensive Statistical Methods. Chapman and Hall, London. find
  16. Liang Huang and David Chiang. 2005. Better k-best parsing. In Proceedings of the 9th International Workshop on Parsing Technologies (IWPT 2005), pages 53–64, Vancouver, Canada. find
  17. Richard Johansson and Alessandro Moschitti. 2010. Syntactic and semantic structure for opinion expression detection. In Proceedings of the Fourteenth Conference on Computational Natural Language Learning, pages 67– 76. find
  18. Dan Klein and Christopher D. Manning. 2003. Fast exact inference with a factored model for natural language parsing. Advances in Neural Information Processing Systems 15 (NIPS 2002), Cambridge, MA: MIT Press. find
  19. John Lafferty, Andrew McCallum, and Fernando Pereira. 2001. Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In 18th International Conf. on Machine Learning. Morgan Kaufmann. find
  20. Mirella Lapata. 2003. Probabilistic text structuring: Experiments with sentence ordering. In Proceedings of the 41st Meeting of the Association of Computational Linguistics, pages 545–552. find
  21. Mitchell P. Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a Large Annotated Corpus of English: the Penn Treebank. Computational Linguistics, 19(2):313– 330. find our card
  22. Guido Minnen, John Carroll, and Darren Pearce. 2001. Applied morphological processing of English. Natural Language Engineering. find
  23. Emily Pitler and Ani Nenkova. 2009. Using syntax to disambiguate explicit discourse connectives in text. In Proceedings of the 47 th Annual Meeting of the Association for Computational Linguistics and the 4 th International Joint Conference on Natural Language Processing. find our card
  24. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi, and Bonnie Webber. 2008. The Penn Discourse Treebank 2.0. In Proceedings of the 6 th International Conference on Languages Resources and Evaluations (LREC 2008), Marrakech, Morocco. find our card
  25. Robert E. Schapire and Yoav Freund. 1999. Large margin classification using the perceptron algorithm. Machine Learning Journal, 37(3):277–296. find
  26. Libin Shen and Aravind Joshi. 2003. An svm based voting algorithm with application to parse reranking. In CoNLL 2003. find
  27. Radu Soricut, Daniel Marcu, et al. 2003. Sentence level discourse parsing using syntactic and lexical information. In Proceedings of the Human Language Technology and North American Association for Computational Linguistics Conference (HLT/NAACL), May 27-June 1. find our card
  28. Radu Soricut, Daniel Marcu, et al. 2006. Stochastic coherence modeling, parameter estimation and decoding for text planning applications. In Proceedings of ACL-2006 (Poster), pages 803–810. find
  29. Kristina Toutanova, Aria Haghighi, Christopher D. Manning, et al. 2008. Kristina toutanova, aria haghighi, and christopher d. manning, a global joint model for semantic role labeling. Computational Linguistics. find
  30. V. Vapnik. 1995. The Nature of Statistical Learning Theory. Springer-Verlag. find
  31. Eric W. Weisstein. 2012. “significant digits.” from mathworld–a wolfram web resource. find
  32. Hiroyasu Yamada and Yuji Matsumoto. 2003. Statistical dependency analysis with support vector machines. In Proceedings of 8th International Workshop on Parsing Technologies. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/ghosh-2012-global-features-shallow-discourse-parsing/