A Dependency Treebank for English

What it is about

The paper describes the syntactic layer of the ISLE project, a small corpus of about 13,000 words of transcribed travel-agent speech with multiple annotation levels. The authors choose dependency rather than phrase structure because predicate-argument structure is what applications need, and because phrase-structure treebanks such as the Penn Treebank need head-percolation heuristics to get it. The annotation is a single monostratal dependency level with no separate surface tree; function words attach to their lexical heads, so auxiliaries depend on the main verb. Each word carries a surface role and a deep role. On one 272-word dialogue, inter-annotator agreement is 93 to 94% for POS and features, 95% for the surface role, and at least 99% for the other fields. The paper reports a pace of about 30 words per hour.

Ideas

• I1. Per-field agreement report: adopt the paper's field-by-field agreement table (POS, features, role) for the annot silver checks, computed on 300 EWT sentences between Haiku and Opus output, so that the weak fields are visible per batch.
• I2. Spoken-dialogue probe: report LAS on the ATIS queries (current 63.0, lowercased) split by whether the sentence contains an auxiliary or a fronted wh-word, since these are the constructions the paper works through.
• I3. Dialogue test set (no training): build a held-out set of short dialogue turns in treebanks/en/ from material we are free to use, to test en on speech-like input; keep it out of the training pool.

For Mova

The code is annot (batch annotation and its gates) and en parse-eval for the probe. The chapters are 26. Questions (wh-questions and inversion) and 37-register-and-variation.md (speech-like registers). The paper's main lesson for Mova is that per-field agreement, not one overall score, decides where the teacher is trusted.

Topics: Treebanks

Cite

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

@inproceedings{rambow2002dependency,
  title = {{A Dependency Treebank for English}},
  author = {Rambow, Owen and Creswell, Cassandre and Szekely, Rachel and Taber, Harriet and Walker, Marilyn},
  booktitle = {Proceedings of the Third International Conference on Language Resources and Evaluation (LREC'02)},
  year = {2002},
  url = {https://aclanthology.org/L02-1325/}
}

References

14 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Srinivas Bangalore and Aravind Joshi. 1999. Supertagging: An approach to almost parsing. Computational Linguistics, 25(2):237–266. find
2. Srinivas Bangalore and Owen Rambow. 2000. Exploiting a probabilistic hierarchical model for generation. In Proceedings of the 18th International Conference on Computational Linguistics (COLING 2000), Saarbrücken, Germany. find
3. Alena Böhmová, Jan Hajič, Eva Hajičová, and Barbora Hladká. 2001. The Prague Dependency Treebank: Three-Level Annotation Scenario. In Anne Abeillé, editor, Treebanks: Building and Using Syntactically Annotated Corpora. Kluwer Academic Publishers. find
4. Michael Collins. 1997. Three generative, lexicalised models for statistical parsing. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics, Madrid, Spain, July. find
5. Jan Hajic, Eva Hajicová, Martin Holub, Petr Pajas, Petr Sgall, Barbora Vidová-Hladká, and Veronika eznícková. 2001. The current status of the prague dependency treebank. In LNAI 2166, LNAI 2166, pages 11–20. Springer Verlag, Berlin, Heidelberg, New York. find
6. Aravind K. Joshi and Yves Schabes. 1991. Tree-adjoining grammars and lexicalized grammars. In Maurice Nivat and Andreas Podelski, editors, Definability and Recognizability of Sets of Trees. Elsevier. find
7. Ronald M. Kaplan and Joan W. Bresnan. 1982. Lexical-functional grammar: A formal system for grammatical representation. In J. W. Bresnan, editor, The Mental Representation of Grammatical Relations. MIT Press, Cambridge, Mass., December. find
8. David Magerman. 1995. Statistical decision-tree models for parsing. In 33rd Meeting of the Association for Computational Linguistics (ACL’95). find
9. Mitchell M. Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a Large Annotated Corpus of English: The Penn Treebank. Computational Linguistics, 19.2:313–330, June. find our card
10. M. Marcus, G. Kim, M. Marcinkiewicz, R. MacIntyre, A. Bies, M. Ferguson, K. Katz, and B. Schasberger. 1994. The penn treebank: Annotating predicate argument structure. In Proceedings of the ARPA Human Language Technology Workshop. find our card
11. Igor A. Mel’čuk. 1988. Dependency Syntax: Theory and Practice. State University of New York Press, New York. find
12. Carl Pollard and Ivan Sag. 1994. Head-Driven Phrase Structure Grammar. University of Chicago Press, Chicago. find
13. P. Sgall, E. Hajicova, and J. Panevova. 1986. The meaning of the sentence and its semantic and pragmatic aspects. Reidel, Dordrecht. find
14. Fei Xia and Martha Palmer. 2001. Converting dependency structure to phrase structures. In hlt2001, pages 61–65. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/bank-2002-rambow-dependency-treebank-english/