Parser combinators for Tigrinya and Oromo morphology

What it is about

The authors write rule-based morphological parsers for Tigrinya and Oromo as parser combinators in Python, not as finite-state transducers. The work was done within a few days during the LoReHLT17 surprise-language evaluation. One grammar yields several outputs at once: breakdown, lemma, gloss and a natural English gloss, ranked by penalties. Tigrinya internal plurals are handled by regular-expression templates over consonant classes, and a Cost channel penalises over-analysis, for example place names that end in -a. The cost of finite-state efficiency is accepted. The evaluation is downstream: adding lemmatization raised entity detection and linking (typed_mention_ceaf_plus, Tigrinya 0.478 to 0.521, Oromo 0.355 to 0.376) and situation-frame F1 (Tigrinya 0.325 to 0.333, Oromo 0.047 to 0.086).

Ideas

• I1. A cost channel for segmentation in tools/hbo: each prefix or suffix cut carries a penalty, so an unsplit word wins unless the stem is in the lexicon. This stops over-segmentation of names, the same way the paper stops ʔuɡanɨda from being read as a possessive -a. Where: the ranking of the lex and pre splits.
• I2. Templates as explicit patterns for Hebrew and Aramaic root-and-pattern plurals: a consonant-class pattern over the root triple, applied after the lexicon lookup, as the paper does for Tigrinya internal plurals. Where: The Noun, the Adjective and the Numeral, The noun in Biblical Aramaic, implemented as a fallback in tools/hbo.
• I3. Measure every added rule as the paper does, by its downstream effect (here, the lemma versus the word), with a before/after table. A rule is kept only if the downstream number rises on held-out data. Where: 13. Biblical Hebrew in Mova: the tools.

For Mova

I1 goes into tools/hbo (segmenter ranking). I2 goes into The Noun, the Adjective and the Numeral and The noun in Biblical Aramaic, with its code in tools/hbo. I3 is the evaluation rule for 13. Biblical Hebrew in Mova: the tools.

Cite

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

@inproceedings{littell2018parser,
  title = {{Parser combinators for Tigrinya and Oromo morphology}},
  author = {Littell, Patrick and McCoy, Tom and Han, Na-Rae and Rijhwani, Shruti and Sheikh, Zaid and Mortensen, David and Mitamura, Teruko and Levin, Lori},
  booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)},
  year = {2018},
  url = {https://aclanthology.org/L18-1611/}
}

References

16 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Bower, D., Arppe, A., Lachler, J., Moshagen, S., and Trosterud, T. (2017). A morphological parser for Odawa. In Proceedings of 2nd Workshop on Computational Methods for Endangered Languages (ComputEL-2). find
2. Burge, W. H. (1975). Recursive Programming Techniques. Addison Wesley. find
3. Frost, R. and Launchbury, J. (1989). Constructing natural language interpreters in a lazy functional language. The Computer Journal, 32:108–121. find
4. Gasser, M. (2011). HornMorpho: a system for morphological processing of Amharic, Oromo, and Tigrinya. In Proceedings of the Conference on Human Language Technology for Development, Alexandria, Egypt. find
5. Hulden, M. (2009). Foma: A finite-state compiler and library. In Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics, pages 29–32. Association for Computational Linguistics. find
6. Hutton, G. and Meijer, E. (1988). Monadic parser combinators. Journal of Functional Programming, 8:437–444. find
7. Hutton, G. (1992). Higher-order functions for parsing. Journal of Functional Programming, 2:323–343, July. find
8. Karttunen, L. and Beesley, K. R. (1992). Two-Level Rule Compiler. Technical Report ISTL-1992-2. Xerox Palo Alto Research Center, Palo Alto, CA. find
9. Karttunen, L. (1993). Finite-state lexicon compiler. Technical Report ISTL-NLTT-1993-04-02. Xerox Palo Alto Research Center, Palo Alto, CA. find
10. Littell, P., Price, K., and Levin, L. (2014). Morphological parsing of Swahili using crowdsourced lexical resources. In Nicoletta Calzolari, et al., editors, Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 3333–3339, Reykjavik, Iceland, May. European Language Resources Association (ELRA). ACL Anthology Identifier: L14-1686. ↗
11. Oard, D. W. (2003). The surprise language exercises. ACM Transactions on Asian Language Information Processing (TALIP), 2(2):79–84, June. find
12. Pan, X., Zhang, B., May, J., Nothman, J., Knight, K., and Ji, H. (2017). Cross-lingual name tagging and linking for 282 languages. In ACL. find
13. Pretorius, L. and Bosch, S. E. (2003). Finite-state computational morphology: An analyzer prototype for Zulu. Machine Translation, pages 191–212. find
14. Strassel, S. and Tracey, J. (2016). LORELEI language packs: Data, tools, and resources for technology development in low resource languages. In LREC 2016: 10th Edition of the Language Resources and Evaluation Conference, Portoroz, pages 3273–3280. find
15. Strassel, S., Bies, A., and Tracey, J. (2017). Situational awareness for low resource languages: the LORELEI situation frame annotation task. In SMERP2017: First International Workshop on Exploitation of Social Media for Emergency Relief and Preparedness, Aberdeen. find
16. Wadler, P. (1985). How to replace failure by a list of successes: A method for exception handling, backtracking, and pattern matching in lazy functional languages. In Conference on Functional Programming Languages and Computer Architecture, pages 113–128. Springer. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/lingbulk-2018-littell-parser-combinators-tigrinya-oromo-morphology/