Extended Constituent-to-Dependency Conversion for English

What it is about

The paper argues that the usual head-percolation conversion of the Penn Treebank into dependencies throws away information that the newer Treebank versions contain. It builds a converter that uses the function tags and the empty-category co-indexation of Penn Treebank II to add richer edge labels and long-distance links for wh-movement, topicalization, passives, clefts and gapping. About 6 percent of the resulting trees are non-projective, and two statistical parsers lose accuracy on the new format (labelled accuracy falls from 90.30 to 87.63 for one parser and from 92.04 to 86.92 for the other). The richer labels pay off in a FrameNet role classifier, whose accuracy rises from 64.3 to 72.5 percent, a 23 percent error reduction.

Ideas

• I1. Compare our converter with the paper: check the head rules and the secondary-edge set (its Tables 1 and 2) against en/src/convert.rs, and list the edge types the mova dialect keeps; measure edge recall on PTB section 23 (test only, LDC data, never shipped).
• I2. Per-edge precision and recall as in the paper's Table 4, for the edge types our converter adds (gap, expletive, cleft), on PTB test only, before any parser is trained on the converted trees.
• I3. Finer role labels: in prag, test whether separating direct object, indirect object and the by-phrase raises role accuracy, on a small set of our own sentences (not FrameNet data), so that the test is ours to publish.

For Mova

The converter is en/src/convert.rs, the dialect is in The Mova dialect: a UD metadialect, treebanks, and converters, and the role side is prag and world. The chapter is 21. The simple clause, and 28. Voice and the passive for the by-phrase. Measure first the edge recall on PTB section 23 for the long-distance edges, which is where the paper's extra information lies.

Topics: Treebanks

Cite

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

@inproceedings{johansson2007extended,
  title = {{Extended Constituent-to-Dependency Conversion for English}},
  author = {Johansson, Richard and Nugues, Pierre},
  booktitle = {Proceedings of the 16th Nordic Conference of Computational Linguistics (NODALIDA 2007)},
  year = {2007},
  url = {https://aclanthology.org/W07-2416/}
}

References

13 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. The Berkeley FrameNet Project. In Proceedings of COLING-ACL’98. find our card
2. Michael J. Collins. 1999. Head-driven statistical models for natural language parsing. Ph.D. thesis, University of Pennsylvania, Philadelphia. find our card
3. David M. Magerman. 1994. Natural language parsing as statistical pattern recognition. Ph.D. thesis, Stanford University. find
4. 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. ↗ our card
5. Mitch Marcus, Grace Kim, Mary Ann Marcinkiewicz, Robert MacIntyre, Ann Bies, Mark Ferguson, Karen Katz, and Britta Schasberger. 1994. The Penn treebank: Annotating predicate argument structure. In ARPA Human Language Technology Workshop. find our card
6. Ryan McDonald and Fernando Pereira. 2006. Online learning of approximate dependency parsing algorithms. In Proceedings EACL-2006. find our card
7. Igor A. Mel’čuk. 1988. Dependency Syntax: Theory and Practice. State University Press of New York, Albany. find
8. Joakim Nivre, Johan Hall, and Jens Nilsson. 2006. Malt-Parser: A data-driven parser generator for dependency parsing. In Proceedings of LREC2006. find
9. Joakim Nivre. 2006. Inductive Dependency Parsing. Springer Verlag. find
10. Owen Rambow, Cassandre Creswell, Rachel Szekely, Harriet Tauber, and Marilyn Walker. 2002. A dependency treebank for English. In Proceedings of LREC2002. find our card
11. Adwait Ratnaparkhi. 1996. A maximum entropy part-of-speech tagger. In Proceedings of EMNLP-1996. find
12. Matthias Trautner Kromann. 2003. The Danish Dependency Treebank and the DTAG treebank tool. In Proceedings of the Second Workshop on Treebanks and Linguistic Theories (TLT). find
13. 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

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/map-2007-johansson-extended-constituent-dependency-conversion-english/