From TreeBank to PropBank

What it is about

The paper describes the first stage of PropBank, which adds predicate-argument labels to the verb instances of the Penn Treebank. Arguments are numbered Arg0 to Arg5 with a function tag for adjuncts (temporal, locative, manner and others), and the numbering is per verb sense, divided by syntax rather than by fine semantics. Frame files give each verb's syntactic realisations and argument labels with examples; roles are kept consistent across near-synonyms and Levin classes, and "metaframes" such as the exchange frame group buy, sell and purchase. By April 2002 frames existed for over 850 verbs, with 30 to 40 added each week, and copying frames from a verb to its Levin-class members needed correction for only one of 13 generated frames. Annotators handled about 50 sentences an hour, and inter-annotator agreement by tagged constituent ran between 80 and 100 percent.

Ideas

• I1. Frame files for the school verbs: one file per verb with numbered roles (agent, theme, source, benefactive), kept in en/src/dict.rs or world/src/dictionary.rs and mapped to the world's roles. Start with the verbs in the first school lessons.
• I2. Copy the roleset of a verb to its class members, as the paper does with destroy, and count how many generated frames need correction. The paper's rate of one in 13 gives a reference point.
• I3. Metaframes for school stories: group buy, sell and pay into an exchange frame in world/src/story.rs, so that a story about shopping answers questions about any of the three verbs.
• I4. Map the functional tags (temporal, locative, manner, cause) onto the adjunct roles in prag/src/roles.rs, and check the mapping against the UD relations of adverbial modifiers on a sample.

For Mova

The chapters are 13. Verbs (verb frames) and 28. Voice and the passive (alternations that keep the same role). The code is en/src/dict.rs, world/src/dictionary.rs and prag/src/roles.rs. The measures are the frame coverage of the school verbs, the copy-correction rate, and the agreement of the adjunct mapping on a sample.

Topics: Treebanks · Semantic role labeling

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.63317/3gh6v2hd4z55

@inproceedings{kingsbury2002treebank,
  title = {{From TreeBank to PropBank}},
  author = {Kingsbury, Paul and Palmer, Martha},
  booktitle = {Proceedings of the Third International Conference on Language Resources and Evaluation (LREC'02)},
  year = {2002},
  doi = {10.63317/3gh6v2hd4z55},
  url = {https://doi.org/10.63317/3gh6v2hd4z55}
}

References

13 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Baker, C.F., Fillmore, C.J., & Lowe, J.B. (1998). The Berkeley Framenet Project. In Proceedings of COLING/ACL (pp 86−90). Montreal: Association for Computational Linguistics. find our card
2. Charniak, E. (1995). Parsing with Context−Free Grammars and Word Statistics. In Technical Report: CS−95−28, Brown University. find
3. Collins, M. (1997). Three generative, lexicalised models for statistical parsing. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics. (pp 16−23) Madrid, Spain: Association for Computational Linguistics. ↗
4. Collins, M. (2000). Discriminative reranking for natural language parsing. In International Conference on Machine Learning. Morgan Kaufmann. (pp 175−182) San Francisco: Dang, H.T., Kipper, K., Palmer, M., & Rosenzweig, J. (1998) Investigating Regular Sense Extensions based on Intersective Levin Classes. In Proceedings of Coling−ACL98. Montreal: Association for Computational Linguistics. find our card
5. Gildea, D. (2001). Statistical Language Understanding Using Frame Semantics. Ph.D. thesis, University of California at Berkeley, Department of Computer and Information Sciences. find
6. Hajicova, E., Panevova, J., & Sgall, P. (2001) Tectogrammatics in Corpus Tagging. In Perspectives on Semantics, Pragmatics, and Discourse: A Festschrift for Ferenc Keifer, I. Kenesei and R.M. Harnish eds. (pp 294−299) Amsterdam/Philadelphia: John Benjamins Publishing Co. ↗
7. Han, C., Lavoie B., Palmer, M., Rambow, O., Kittredge, R., Korelsky, T., Kim, N. & Kim, M. (2000) Handling Structural Divergences and Recovering Dropped Arguments in a Korean/English Machine Translation System. In Proceedings of the Association for Machine Translation in the Americas 2000. (pp 40−53) Berlin/New York: Springer Verlag. find
8. Kingsbury, P., Marcus, M. & Palmer, M. (forthcoming). Adding Predicate Argument Structure to the Penn TreeBank. In Proceedings of the Human Language Technology Conference, San Diego, CA, March 2002. find
9. Kittredge, R., Korelsky, T., Lavoie, B., Palmer, M., Han, C., Park, C., & Bies, A. (2001) SBIR−II: Korean− English Machine Translation of Battlefield Messages. Final Report to the Army Research Lab. find
10. Levin, B. (1993) English Verb Classes and Alternations A Preliminary Investigation. Chicago: University of Chicago Press. find
11. Marcus, M. (1994). The Penn treebank: A revised corpus design for extracting predicate−argument structure. In Proceedings of the ARPA Human Language Technology Workshop, Princeton, NJ. find
12. Marcus, M., Santorini, B., & Marcinkiewicz, M.A. (1993). Building a large annotated corpus of English : the Penn Treebank. Computational linguistics 19(2), 313−330 ↗ our card
13. Miller, G., Beckwith, R., Fellbaum, C., Gross, D., & Miller, K. (1993) Five papers on wordnet. International Journal of Lexicography 3(4), 235−312. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/kingsbury-2002-treebank-propbank/