Morphological Features for Parsing Morphologically-rich Languages: A Case of Arabic

What it is about

Dehdari, Tounsi and van Genabith ask which morphological features, folded into the POS tag, help a PCFG latent-annotation parser (a re-implemented Berkeley parser) on the Penn Arabic Treebank, using the ATB3-v3.2 80/10/10 split. Of the 498 full ATB tags, they identify 15 features that can be included or excluded, giving 32,768 combinations, too many to try all. They compare non-iterative search, greedy search, a merged heuristic, best-first search with backtracking, and simulated annealing (mean dev F 82.84, standard deviation 0.21 over 50 runs). Backtracking finds the best set: genitive clitics, proper noun, negation and determiner. The conclusion reports F 85.03 for vocalized and 83.34 for unvocalized no-tag parsing, with product models and unknown-word signature classes. Case helped gold parsing but not predicted parsing, and the authors conclude that no tagset is best for every task.

Ideas

• I1. A deterministic best-first search with backtracking over Mova's feature sets for gparse, with ordered sets (BTreeSet) and a two-run determinism check. Skip simulated annealing, which is random and costly.
• I2. Unknown-word signature classes: map unseen tokens to a small set of classes by prefix, suffix and shape heuristics. For Hebrew this could use the common prefix letters, the suffixes ים and ות, and the noun shapes from the noun chapter. Measure it on held-out unseen tokens.
• I3. The merged heuristic M(f) = (S_top − S_not-f) + (S_f − S_bottom): a cheap way to rank features from two single-feature passes, used to order the trials of I1.

For Mova

I1 is an experiment plan for tools/hbo and gparse, to be run on dev before any claim. I2 goes into the unknown-word handling of tools/hbo (unpointed mode) and into The Noun, the Adjective and the Numeral. I3 is the ordering rule for the search in I1.

Cite

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

@inproceedings{dehdari2011morphological,
  title = {{Morphological Features for Parsing Morphologically-rich Languages: A Case of Arabic}},
  author = {Dehdari, Jon and Tounsi, Lamia and Genabith, Josef van},
  booktitle = {Proceedings of the Second Workshop on Statistical Parsing of Morphologically Rich Languages},
  year = {2011},
  url = {https://aclanthology.org/W11-3802/}
}

References

29 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Attia, M., Foster, J., Hogan, D., Le Roux, J., Tounsi, L., and van Genabith, J. (2010). Handling unknown words in statistical latent-variable parsing models for Arabic, English and French. In NAACL/ HLT Workshop on Statistical Parsing of Morphologically Rich Languages, pages 67–75. find
2. Bengoetxea, K. and Gojenola, K. (2010). Application of different techniques to dependency parsing of Basque. In NAACL/ HLT Workshop on Statistical Parsing of Morphologically Rich Languages, pages 37–39. find
3. Bikel, D. (2004). On the Parameter Space of Generative Lexicalized Parsing Models. PhD thesis, University of Pennslyvania. find
4. Collins, M., Hajič, J., Ranshaw, L., and Tillman, C. (1999). A statistical parser for Czech. In Proceedings of the Annual Meeting of the Association for Computational Linguistics. 9 NP’s comprise 58% of all non-terminals in the treebank that we used. They are more than five times more frequent than the next most common non-terminal (PP). find
5. Cowan, B. and Collins, M. (2005). Morphology and reranking for the statistical parsing of Spanish. In Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing, pages 795–802, Vancouver, British Columbia, Canada. Association for Computational Linguistics. find
6. Diab, M. T. (2007). Improved Arabic base phrase chunking with a new enriched POS tag set. In ACL Workshop on Computational Approaches to Semitic Languages: Common Issues and Resources, pages 89–96. find
7. Habash, N. and Rambow, O. (2005). Arabic tokenization, part-of-speech tagging and morphological disambiguation in one fell swoop. In Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics, ACL ’05, pages 573–580, Stroudsburg, PA, USA. Association for Computational Linguistics. find our card
8. Habash, N. and Roth, R. (2009). CATiB: The Columbia Arabic Treebank. In Proceedings of the ACL-IJCNLP 2009 Conference Short Papers, pages 221–224, Suntec, Singapore. Association for Computational Linguistics. find our card
9. Johnson, M. (1998). PCFG models of linguistic tree representations. Computational Linguistics, 24(4):613–632. find
10. Kirkpatrick, S., Gelatt, C. D. J., and Vecchi, M. P. (1983). Optimization by simulated annealing. Science, 220(4598):671–680. find
11. Klein, D. and Manning, C. (2003a). A* parsing: fast exact viterbi parse selection. In Proceedings of the North American Association for Association for Computational Linguistics. find
12. Klein, D. and Manning, C. (2003b). Accurate unlexicalized parsing. In Proceedings of the 41st Annual Meeting of the ACL, pages 423–430. find our card
13. Kulick, S., Gabbard, R., and Marcus, M. (2006). Parsing the Arabic Treebank: Analysis and improvements. In Proceedings of the Treebanks and Linguistic Theories Conference (TLT-2006), pages 31–42, Prague, Czech Republic. find
14. Maamouri, M., Bies, A., and Kulic, S. (2009). Creating a methodology for large-scale correction of treebank annotation: The case of the Arabic treebank. In Choukri, K. and Maegaard, B., editors, MEDAR Second International Conference on Arabic Language Resources and Tools, Egypt. find
15. Marcus, M., Kim, G., Marcinkiewicz, M. A., Mac-Intyre, R., Bies, A., Ferguson, M., Katz, K., and Schasberger, B. (1994). The Penn Treebank: Annotating predicate argument structure. In Proceedings of the 1994 ARPA Speech and Natural Language Workshop, pages 114–119, Princeton, NJ, USA. find our card
16. Marton, Y., Habash, N., and Rambow, O. (2010). Improving Arabic dependency parsing with lexical and inflectional morphological features. In Proceedings of the NAACL HLT 2010 First Workshop on Statistical Parsing of Morphologically-Rich Languages, pages 13–21, Los Angeles, CA, USA. Association for Computational Linguistics. find our card
17. Marton, Y., Habash, N., and Rambow, O. (2011). Improving Arabic dependency parsing with form-based and functional morphological features. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pages 1586–1596, Portland, Oregon, USA. Association for Computational Linguistics. find our card
18. Matsuzaki, T., Miyao, Y., and Tsujii, J. (2005). Probabilistic CFG with latent annotations. In Proceedings of the 43rd Annual Meeting of the ACL, pages 75–82, Ann Arbor, MI, USA. find
19. McDonald, R. and Crammer, K. (2005). Online large-margin training of dependency parsers. In Proceedings of the Annual Meeting of the Association for Computational Linguistics. find our card
20. Nivre, J., Hall, J., Kübler, S., McDonald, R., Nilsson, J., Riedel, S., and Yuret, D. (2007a). find
21. The CoNLL 2007 shared task on dependency parsing. In Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL 2007, pages 915– 932, Prague, Czech Republic. Association for Computational Linguistics. find
22. Nivre, J., Hall, J., Nilsson, J., Chanev, A., Eryigit, G., Kübler, S., Marinov, S., and Marsi, E. (2007b). Maltparser: A language-independent system for data-driven dependency parsing. Natural Language Engineering, 13(2):95–135. find
23. Petrov, S. (2009). Coarse-to-Fine Natural Language Processing. PhD thesis, University of California at Berkeley, Berkeley, CA, USA. find
24. Petrov, S. (2010). Products of random latent variable grammars. In Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), pages 19–27, Los Angeles, CA, USA. Association for Computational Linguistics. find
25. Petrov, S., Barrett, L., Thibaux, R., and Klein, D. (2006). Learning accurate, compact and interpretable tree annotation. In Proceedings of the 21st International Conference on Computational Linguistics and the 44th Annual Meeting of the ACL, Sydney, Australia. find our card
26. Petrov, S. and Klein, D. (2007). Improved inference for unlexicalized parsing. In Proceedings of HLT-NAACL, Rochester, NY, USA. find
27. Satoshi, S. and Collins, M. (1997). Evalb bracket scoring program. find
28. Tsarfaty, R. and Sima’an, K. (2008). Relational-realizational parsing. In Proceedings of Conference on Computational Linguistics (CoLing), pages 18–22. find
29. Zhou, R. and Hansen, E. A. (2005). Beam-stack search: Integrating backtracking with beam search. In Proceedings of the International Conference on Automated Planning and Scheduling (ICAPS), pages 90–98, Monterey, CA, USA. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2011-dehdari-morphological-features-parsing-morphologically-rich/