Elamur Research⌕
☰
home / papers / 2020 / UniMorph 3.0: Universal Morphology

UniMorph 3.0: Universal Morphology

Summary

UniMorph provides inflectional paradigms for many languages. Each entry is a triple of lemma, feature bundle and inflected form, tagged with one cross-lingual feature schema. Release 3.0 responds to community criticism of version 2.0. An audit by Gorman et al. (2019) found that many sampled entries were mis-scraped or already wrong in Wiktionary, and Kementchedjhieva et al. (2019) reported coverage too thin to help a Bulgarian lexicon-induction task. The authors rebuilt the Wiktionary extraction so that it keeps only real inflection tables and ties each one to the correct part of speech and lemma. It also collects morphology stored outside tables, such as inherent noun gender. Additional data come from converting existing resources, redistributed with permission. The schema was revised, for example to handle syntactic blends and argument politeness. Three tools were released: two validators for data creators and one command-line utility for looking up paradigms and analyses. The resource now covers 118 languages and 14.8 million triples. This matters because UniMorph is the data behind several SIGMORPHON shared tasks, and cleaner, fuller morphology benefits parsing, translation and cross-lingual transfer.

Key points

  • UniMorph 3.0 adds 66 languages, reaching 118 languages and 14.8 million (lemma, features, inflection) triples. For comparison, 2.0 covered 52 languages.
  • The data span 16 language families plus 2 isolates. The paper reports new parts of speech for 12 languages in one place and new word classes for 85 languages in another.
  • Gorman et al. (2019) judged about 25% of a sample from the CoNLL–SIGMORPHON 2017 data to be erroneous. The worst systematic problems were Romanian adjectives, Hungarian and Finnish nouns, and Latin lemmas.
  • The new table isolation sharply reduced German layouts across nouns, verbs and adjectives, from 171 shapes to 14. The most frequent noun layout grew from 9,323 to 14,181 occurrences.
  • Extraction now also captures determiners, numerals, pronouns and proper nouns. It also fixed a bug that merged data across languages whose names end the same way, such as Ancient Greek leaking into Greek.

Interesting ideas

  • I1. Use markup the page already has. Instead of guessing which HTML tables matter, the scraper relies on the class="inflection-table" attribute and on class-tagged headings for POS and lemma. Machine-generated structural markup is often a more reliable filter than layout heuristics when scraping any wiki-like source.
  • I2. Annotate one template, apply it to all. Annotators label cells in a single sample table per layout, and the mapping is propagated to every lemma sharing that shape. Reducing layout variety first, as the German example shows, raises the payoff per hour of human work and lowers error rates.
  • I3. Store inherent lexical features with inflections. Adding properties like noun gender acknowledges that some inflection patterns cannot be predicted from the lemma's form alone. Any dataset for inflection or generation tasks should check whether the task is fully specified without such features.
  • I4. Keep the lemma separate from the page title. Treating headword and lemma as distinct preserves distinctions like the Latin macrons in malus vs. mālus. The same issue arises whenever dictionary page keys are normalized more than the actual entries.
  • I5. Give contributors validation tools. Shipping validators to resource creators catches schema violations before release. This is a cheap way to keep a collaboratively built multilingual resource consistent.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Cited by

68 papers of our library cite it (of the 140673 whose bibliographies we read); 10 of them have a page here, 58 not yet.

Cite

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

@inproceedings{mccarthy2020unimorph,
  title = {{UniMorph 3.0: Universal Morphology}},
  author = {McCarthy, Arya D. and Kirov, Christo and Grella, Matteo and Nidhi, Amrit and Xia, Patrick and Gorman, Kyle and Vylomova, Ekaterina and Mielke, Sabrina J. and Nicolai, Garrett and Silfverberg, Miikka and Arkhangelskiy, Timofey and Krizhanovsky, Nataly and Krizhanovsky, Andrew and Klyachko, Elena and Sorokin, Alexey and Mansfield, John and Ernštreits, Valts and Pinter, Yuval and Jacobs, Cassandra L. and Cotterell, Ryan and Hulden, Mans and Yarowsky, David},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  doi = {10.3929/ethz-b-000462327},
  url = {https://doi.org/10.3929/ethz-b-000462327}
}

References

49 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abeillé, A., Clément, L., and Toussenel, F., (2003). Building a Treebank for French, pages 165–187. Springer Netherlands, Dordrecht. find
  2. Adams, O., Wiesner, M., Trmal, J., Nicolai, G., and Yarowsky, D. (2019). Induced Inflection-Set Keyword Search in Speech. arXiv e-prints, page arXiv:1910.12299, Oct. find
  3. Adelaar, K. A. and Himmelmann, N. (2005). The Austronesian languages of Asia and Madagascar, volume 7. Psychology Press. find
  4. Agić, Ž. and Vulić, I. (2019). JW300: A wide-coverage parallel corpus for low-resource languages. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3204–3210, Florence, Italy, July. Association for Computational Linguistics. find our card
  5. Arkhangelskiy, T. and Medvedeva, M. (2016). Developing morphologically annotated corpora for minority languages of russia. In CLiF. find
  6. Arkhangelskiy, T., Belyaev, O., and Vydrin, A. (2012). The creation of large-scale annotated corpora of minority languages using UniParser and the EANC platform. In Proceedings of COLING 2012: Posters, pages 83–92, Mumbai, India, December. find
  7. The COLING 2012 Organizing Committee. find
  8. Blevins, T. and Zettlemoyer, L. (2019). Better character language modeling through morphology. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1606–1613, Florence, Italy, July. Association for Computational Linguistics. find
  9. Brykina, M., Orlova, S., and Wagner-Nagy, B. (2018). INEL Selkup Corpus. Version 0.1. In The INEL corpora of indigenous Northern Eurasian languages., Hamburg, December. Hamburger Zentrum für Sprachkorpora. find
  10. Cotterell, R., Kirov, C., Sylak-Glassman, J., Yarowsky, D., Eisner, J., and Hulden, M. (2016a). find
  11. The SIGMORPHON 2016 shared Task—Morphological reinflection. In Proceedings of the 14th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 10–22, Berlin, Germany, August. Association for Computational Linguistics. find
  12. Cotterell, R., Schütze, H., and Eisner, J. (2016b). Morphological smoothing and extrapolation of word embeddings. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1651–1660, Berlin, Germany, August. Association for Computationa… find
  13. Cotterell, R., Kirov, C., Sylak-Glassman, J., Walther, G., Vylomova, E., Xia, P., Faruqui, M., Kübler, S., Yarowsky, D., Eisner, J., and Hulden, M. (2017). find
  14. CoNLL-SIGMORPHON 2017 shared task: Universal morphological reinflection in 52 languages. In Proceedings of the CoNLL SIGMORPHON 2017 Shared Task: Universal Morphological Reinflection, pages 1–30, Vancouver, August. Association for Computational Linguistics. find
  15. Cotterell, R., Kirov, C., Sylak-Glassman, J., Walther, G., Vylomova, E., McCarthy, A. D., Kann, K., Mielke, S., Nicolai, G., Silfverberg, M., Yarowsky, D., Eisner, J., and Hulden, M. (2018). The CoNLL–SIGMORPHON 2018 shared task: Universal morphological reinflection. In Proceedings of the CoNLL–SIGM… find
  16. Creutz, M., Hirsimäki, T., Kurimo, M., Puurula, A., Pylkkönen, J., Siivola, V., Varjokallio, M., Arisoy, E., Saraçlar, M., and Stolcke, A. (2007). Analysis of morph-based speech recognition and the modeling of out-of-vocabulary words across languages. In Human Language Technologies 2007: The Confere… find
  17. Dyer, C., Muresan, S., and Resnik, P. (2008). Generalizing word lattice translation. In Proceedings of ACL-08: HLT, pages 1012–1020, Columbus, Ohio, June. Association for Computational Linguistics. find
  18. Gorman, K. and Yang, C. (2019). When nobody wins. In Franz Rainer, et al., editors, Competition in Inflection and Word-Formation, pages 169–193. Springer, Dordrecht. find
  19. Gorman, K., McCarthy, A. D., Cotterell, R., Vylomova, E., Silfverberg, M., and Markowska, M. (2019). Weird inflects but OK: Making sense of morphological generation errors. In Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL), pages 140–151, Hong Kong, China, Nove… find
  20. Grella, M. (2014). English dictionary. LINDAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  21. Grella, M. (2018a). Italian content words v3. LIN-DAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  22. Grella, M. (2018b). Italian function words v3. LIN-DAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  23. Grella, M. (2019a). French morphological dictionary 0.1. LINDAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  24. Grella, M. (2019b). German morphological dictionary 0.1. LINDAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  25. Grella, M. (2019c). Spanish morphological dictionary 0.1. LINDAT/CLARIN digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  26. Halle, M. (1973). Prolegomena to a theory of word formation. Linguistic Inquiry, 4(1):3–16. find
  27. Hohensee, M. and Bender, E. M. (2012). Getting more from morphology in multilingual dependency parsing. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 315–326, Montréal, Canada, June. Associatio… find
  28. Kementchedjhieva, Y., Hartmann, M., and Søgaard, A. (2019). Lost in evaluation: Misleading benchmarks for bilingual dictionary induction. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processin… find
  29. Kim, N., Patel, R., Poliak, A., Xia, P., Wang, A., McCoy, T., Tenney, I., Ross, A., Linzen, T., Van Durme, B., Bowman, S. R., and Pavlick, E. (2019). Probing what different NLP tasks teach machines about function word comprehension. In Proceedings of the Eighth Joint Conference on Lexical and Comput… find our card
  30. Kirov, C., Sylak-Glassman, J., Que, R., and Yarowsky, D. (2016). Very-large scale parsing and normalization of Wiktionary morphological paradigms. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 3121–3126, Portorož, Slovenia, May. European L… find our card
  31. Kirov, C., Cotterell, R., Sylak-Glassman, J., Walther, G., Vylomova, E., Xia, P., Faruqui, M., Mielke, S., McCarthy, A., Kübler, S., Yarowsky, D., Eisner, J., and Hulden, M. (2018). UniMorph 2.0: Universal morphology. In Proceedings of the Eleventh International Conference on Language Resources and… find
  32. Klyachko, E., Sorokin, A., Krizhanovskaya, N., Krizhanovsky, A., and Ryazanskaya, G. (2019). LowResourceEval-2019: A shared task on morphological analysis for low-resource languages. In Papers from the Annual International Conference “Dialogue” (2019), Issue 18, pages 45–62, Moscow. Computational Li… find
  33. Lepschy, A. L. and Lepschy, G. C. (1981). La lingua italiana. Bompiani Milan:. find
  34. Makarov, P. and Clematide, S. (2018). UZH at CoNLL– SIGMORPHON 2018 shared task on universal morphological reinflection. In Proceedings of the CoNLL– SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection, pages 69–75, Brussels, October. Association for Computational Linguistics. find
  35. Mansfield, J. (2019). Murrinhpatha morphology and phonology, volume 653. Walter de Gruyter GmbH & Co KG. find
  36. McCarthy, A. D., Silfverberg, M., Cotterell, R., Hulden, M., and Yarowsky, D. (2018). Marrying universal dependencies and universal morphology. In Proceedings of the Second Workshop on Universal Dependencies (UDW 2018), pages 91–101, Brussels, Belgium, November. Association for Computational Linguis… find our card
  37. McCarthy, A. D., Vylomova, E., Wu, S., Malaviya, C., Wolf-Sonkin, L., Nicolai, G., Kirov, C., Silfverberg, M., Mielke, S. J., Heinz, J., Cotterell, R., and Hulden, M. (2019). find
  38. The SIGMORPHON 2019 shared task: Morphological analysis in context and cross-lingual transfer for inflection. In Proceedings of the 16th Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 229–244, Florence, Italy, August. Association for Computational Linguistics. find
  39. Narasimhan, K., Karakos, D., Schwartz, R., Tsakalidis, S., and Barzilay, R. (2014). Morphological segmentation for keyword spotting. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 880–885, Doha, Qatar, October. Association for Computational L… find
  40. Nicolai, G., Lewis, D., McCarthy, A. D., Mueller, A., Wu, W., and Yarowsky, D. (under review). Fine-grained morphosyntactic analysis and generation tools for more than one thousand languages. In Submitted to LREC 2020. find
  41. Nivre, J., de Marneffe, M.-C., Ginter, F., Goldberg, Y., Hajič, J., Manning, C. D., McDonald, R., Petrov, S., Pyysalo, S., Silveira, N., Tsarfaty, R., and Zeman, D. (2016). Universal dependencies v1: A multilingual treebank collection. In Proceedings of the Tenth International Conference on Languag… find our card
  42. Plank, B., Klerke, S., and Agic, Z. (2018). The best of both worlds: Lexical resources to improve low-resource part-of-speech tagging. CoRR, abs/1811.08757. — [arXiv:1811.08757](https://arxiv.org/abs/1811.08757) find
  43. Seeker, W. and Çetinoğlu, Ö. (2015). A graph-based lattice dependency parser for joint morphological segmentation and syntactic analysis. Transactions of the Association for Computational Linguistics, 3:359–373. find our card
  44. Shearing, S., Kirov, C., Khayrallah, H., and Yarowsky, D. (2018). Improving low resource machine translation using morphological glosses (non-archival extended abstract). In Proceedings of the 13th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Papers), pag… find
  45. Sylak-Glassman, J., Kirov, C., Yarowsky, D., and Que, R. (2015). A language-independent feature schema for inflectional morphology. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Vo… find
  46. Tapanainen, P. and Voutilainen, A. (1994). Tagging accurately - don’t guess if you know. In Fourth Conference on Applied Natural Language Processing, pages 47–52, Stuttgart, Germany, October. Association for Computational Linguistics. find
  47. Wicentowski, R. (2002). Modeling and learning multilingual inflectional morphology in a minimally supervised framework. Ph.D. thesis, Johns Hopkins University, Baltimore, Maryland. find
  48. Yarowsky, D. and Ngai, G. (2001). Inducing multilingual POS taggers and NP bracketers via robust projection across aligned corpora. In Second Meeting of the North American Chapter of the Association for Computational Linguistics. find
  49. Zaytseva, N., Krizhanovsky, A., Krizhanovsky, N., Pellinen, N., and Rodionova, A. (2017). Open corpus of Veps and Karelian languages (VepKar): preliminary data collection and dictionaries. In Corpus Linguistics-2017, pages 172– 177. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/mccarthy-2020-unimorph-3-0-universal-morphology/