Elamur Research⌕
☰
home / papers / 2012 / A Universal Part-of-Speech Tagset

A Universal Part-of-Speech Tagset

Summary

Different treebanks use incompatible and highly fine-grained part-of-speech tag sets. This makes it difficult to compare tagging quality across languages and to transfer models from one language to another. The authors propose a compact set of twelve universal part-of-speech categories. For 25 treebanks, they manually created mappings from their tagsets to this set, so that, together with the original corpora, this yields consistently annotated data for 22 languages. The categories were defined practically rather than theoretically: the authors studied the annotation guidelines of each corpus. Contentious cases were resolved pragmatically; for example, Korean stative verbs were assigned to adjectives. The usefulness of the resource is demonstrated in three experiments. The first compares tagging accuracy across all corpora. The second is fully unsupervised grammar induction on automatically projected tags. The third is direct transfer of a delexicalized parser from English to other languages. This simplest approach outperformed the best unsupervised methods of the time. The work unified cross-lingual POS annotation, and the UPOS tag set in Universal Dependencies was later built on it.

Key points

  • Tags: the set contains 12 tags: NOUN, VERB, ADJ, ADV, PRON, DET, ADP, NUM, CONJ, PRT, "." (punctuation), and X (everything else, including abbreviations and foreign words).
  • Mappings: these were created for 25 treebanks in 22 languages, with two corpora each for Chinese, German, and Japanese. The size of the source tagsets ranges from 11 (SynTagRus) to 294 (Sinica). The resource was published at code.google.com/p/universal-pos-tags in early 2011.
  • Tagging accuracy: with universal tags, the spread across languages was halved, from 10.4 to 5.1. The best result is obtained by training on the original tags and then mapping the predictions to universal tags. For Japanese, accuracy exceeds 99%, while Turkish remains the most difficult (90.2%). An English dependency parser using 12 tags instead of the 45 PennTreebank tags loses only about 0.6% accuracy.
  • Grammar induction: the projection tagger of Das & Petrov was combined with the USR model of Naseem et al. The system was evaluated on 8 Indo-European languages from the CoNLL-X corpus, on sentences of up to 10 words. It requires no annotation in the target language, yet it outperforms DMV on all languages and is competitive with PGI.
  • Direct parser transfer: a delexicalized English parser with gold universal tags achieves, for example, 77.9% for Portuguese and 73.9% for Greek (directed dependency accuracy). This is higher than unsupervised grammar induction systems.

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

Topics: Part-of-speech tagging

Cite

The paper itself, not this page. BibTeX from the library's record of the work. arXiv 1104.2086

@inproceedings{petrov2012universal,
  title = {{A Universal Part-of-Speech Tagset}},
  author = {Petrov, Slav and Das, Dipanjan and McDonald, Ryan},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  eprint = {1104.2086},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1104.2086}
}

References

61 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. A. Abeillé, L. Clément, and F. Toussenel. 2003. Building a Treebank for French. In Abeillé (Abeillé, 2003), chapter 10. find
  2. A. Abeillé, editor. 2003. Treebanks: Building and Using Parsed Corpora. Kluwer. find
  3. I. Aduriz, M. J. Aranzabe, J. M. Arriola, A. Atutxa, A. Diaz de Ilarraza, A. Garmendia, and M. Oronoz. 2003. Construction of a Basque dependency treebank. In Proc. of the Workshop on Treebanks and Linguistic Theories. find
  4. S. Afonso, E. Bick, R. Haber, and D. Santos. 2002. Floresta sintá(c)tica: a treebank for Portuguese. In Proc. of LREC. find our card
  5. T. Berg-Kirkpatrick and D. Klein. 2010. Phylogenetic grammar induction. In Proc. of ACL. find
  6. I.M. Boguslavsky, L.L. Iomdin, I.S. Chardin, and L.G. Kreidlin. 2002. Development of a dependency treebank for russian and its possible applications in nlp. In Proc. of LREC. find
  7. A. Böhmová, J. Hajič, E. Hajičová, and B. Hladká. 2003. The PDT: a 3-level annotation scenario. In Abeillé (Abeillé, 2003), chapter 7, pages 103–127. find
  8. S. Brants, S. Dipper, S. Hansen, W. Lezius, and G. Smith. 2002. The TIGER Treebank. In Proc. of the Workshop on Treebanks and Linguistic Theories. find
  9. T. Brants. 1995. Tagset reduction without information loss. Proc. of ACL. find
  10. T. Brants. 2000. TnT - a statistical part-of-speech tagger. In Proc. of ANLP. find our card
  11. S. Buchholz and E. Marsi. 2006. CoNLL-X shared task on multilingual dependency parsing. In Proc. of CoNLL. find our card
  12. A. Carnie. 2002. Syntax: A Generative Introduction (Introducing Linguistics). Blackwell Publishing. find
  13. K. Chen, C. Luo, M. Chang, F. Chen, C. Chen, C. Huang, and Z. Gao. 2003. Sinica treebank: Design criteria, representational issues and implementation. In Abeillé (Abeillé, 2003), chapter 13, pages 231–248. find
  14. C. Christodoulopoulos, S. Goldwater, and M. Steedman. 2010. Two decades of unsupervised POS induction: How far have we come? In Proc. of EMNLP. find
  15. M. Civit and M.A. Martı́. 2004. Building cast3lb: A spanish treebank. Research on Language & Computation, 2(4):549–574. find
  16. S. B. Cohen, D. Das, and N. A. Smith. 2011. Unsupervised structure prediction with non-parallel multilingual guidance. In Proc. of EMNLP. find
  17. D. Csendes, J. Csirik, T. Gyimóthy, and A. Kocsor. 2005. The Szeged Treebank. Springer. find
  18. D. Das and S. Petrov. 2011. Unsupervised part-of-speech tagging with bilingual graph-based projections. In Proc. of ACL-HLT. find our card
  19. J. DeNero and J. Uszkoreit. 2011. Inducing sentence structure from parallel corpora for reordering. In Proc. of EMNLP. find
  20. M. Dickinson and C. Jochim. 2008. A simple method for tagset comparison. Proc. of LREC. find
  21. P. Dienes and C. Oravecz. 2000. Bottom-up tagset design from maximally reduced tagset. In Proc of the COLING Workshop on Linguistically Interpreted Corpora. find
  22. M. Ariel Dominguez and G. Infante-Lopez. 2008. Searching for part of speech tags that improve parsing models. Lecture Notes in Computer Science. find
  23. S. Džeroski, T. Erjavec, N. Ledinek, P. Pajas, Z. Žabokrtsky, and A. Žele. 2006. Towards a Slovene dependency treebank. In Proc. of LREC. find
  24. T. Erjavec. 2004. MULTEXT-East version 3: Multilingual morphosyntactic specifications, lexicons and corpora. In Proc. of LREC. find
  25. N. Evans and S. Levinson. 2009. The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and Brain Sciences, 32(05). find
  26. K. Gimpel, N. Schneider, B. O’Connor, D. Das, D. Mills, J. Eisenstein, M. Heilman, D. Yogatama, J. Flanigan, and N. A. Smith. 2011. Part-of-speech tagging for twitter: Annotation, features, and experiments. In Proc. of ACL. find our card
  27. J. Hajič, O. Smrž, P. Zemánek, J. Šnaidauf, and E. Beška. 2004. Prague Arabic dependency treebank: Development in data and tools. In Proc. of NEMLAR. find
  28. Z. Huang, V. Eidelman, and M. Harper. 2009. Improving simple bigram HMM part-of-speech tagger by latent annotation. In Proc. of NAACL-HLT. find
  29. Y. Kawata and J. Bartels. 2000. Stylebook for the Japanese treebank in VERBMOBIL. find
  30. M.J. Kim. 2002. Does Korean have adjectives? MIT Working Papers in Linguistics, 43:71–89. find
  31. D. Klein and C. D. Manning. 2004. Corpus-based induction of syntactic structure: models of dependency and constituency. In Proc. of ACL. find
  32. M.T. Kromann, L. Mikkelsen, and S.K. Lynge. 2003. Danish Dependency Treebank. In Proc. of the Workshop on Treebanks and Linguistic Theories. find
  33. S. Kurohashi and M. Nagao. 1997. Kyoto University text corpus project. In Proc. of ANLP. find
  34. C. Manning. 2011. Part-of-speech tagging from 97% to 100%: is it time for some linguistics? Proc. of CICLing. find our card
  35. M. P. Marcus, Mary Ann Marcinkiewicz, and Beatrice Santorini. 1993. Building a large annotated corpus of English: the Penn treebank. Computational Linguistics, 19. find our card
  36. M. A. Martı́, M. Taulé, L. Màrquez, and M. Bertran. 2007. CESS-ECE: A multilingual and multilevel annotated corpus. Available for download from: http://www.lsi.upc.edu/∼mbertran/cess-ece/. ↗
  37. R. McDonald and J. Nivre. 2007. Characterizing the errors of data-driven dependency parsing models. In Proc. of EMNLP-CoNLL. find our card
  38. R. McDonald, S. Petrov, and K. Hall. 2011. Multisource transfer of delexicalized dependency parsers. In Proc. of EMNLP. find
  39. S. Montemagni, F. Barsotti, M. Battista, N. Calzolari, O. Corazzari, A. Lenci, A. Zampolli, F. Fanciulli, M. Massetani, R. Raffaelli, R. Basili, M. T. Pazienza, D. Saracino, F. Zanzotto, N. Nana, F. Pianesi, and R. Delmonte. 2003. Building the Italian Syntactic-Semantic Treebank. In Abeillé (Abeill… find
  40. T. Naseem, B. Snyder, J. Eisenstein, and R. Barzilay. 2009. Multilingual part-of-speech tagging: Two unsupervised approaches. JAIR, 36. find
  41. T. Naseem, H. Chen, R. Barzilay, and M. Johnson. 2010. Using universal linguistic knowledge to guide grammar induction. In Proc. of EMNLP. find
  42. F. J. Newmeyer. 2005. Possible and Probable Languages: A Generative Perspective on Linguistic Typology. Oxford University Press. find
  43. J. Nivre, J. Nilsson, and J. Hall. 2006. Talbanken05: A Swedish Treebank with Phrase Structure and Dependency Annotation. In Proc. of LREC. find our card
  44. J. Nivre, J. Hall, S. Kübler, R. McDonald, J. Nilsson, S. Riedel, and D. Yuret. 2007. find
  45. The CoNLL 2007 shared task on dependency parsing. In Proc. EMNLP-CoNLL. find our card
  46. K. Oflazer, B. Say, D. Zeynep Hakkani-Tür, and G. Tür. 2003. Building a Turkish treebank. In Abeillé (Abeillé, 2003), chapter 15, pages 261–277. find
  47. M. Palmer, N. Xue, F. Xia, F. Chiou, Z. Jiang, and M. Chang. 2007. Chinese Treebank 6.0. Technical report, Linguistic Data Consortium, Philadelphia. find
  48. P. Prokopidis, E. Desypri, M. Koutsombogera, H. Papageorgiou, and S. Piperidis. 2005. Theoretical and practical issues in the construction of a Greek dependency treebank. In Proc. of the Workshop on Treebanks and Linguistic Theories. find
  49. O. Rambow, B. Dorr, D. Farwell, R. Green, N. Habash, S. Helmreich, E. Hovy, L. Levin, K. J. Miller, T. Mitamura, Reeder F., and A. Siddharthan. 2006. Parallel syntactic annotation of multiple languages. In Proc. of LREC. find
  50. L. Shen, G. Satta, and A. Joshi. 2007. Guided learning for bidirectional sequence classification. In Proc. of ACL. find
  51. K. Simov, P. Osenova, S. Kolkovska, E. Balabanova, D. Doikoff, K. Ivanova, A. Simov, and M Kouylekov. 2002. Building a Linguistically Interpreted Corpus of Bulgarian: the BulTreeBank. In Proc. of LREC. find
  52. W. Skut, B. Krenn, T. Brants, and H. Uszkoreit. 1997. An annotation scheme for free word order languages. In Proc. of ANLP. find our card
  53. N. A. Smith and J. Eisner. 2005. Contrastive estimation: Training log-linear models on unlabeled data. In Proc. of ACL. find our card
  54. B. Snyder, T. Naseem, J. Eisenstein, and R. Barzilay. 2009. Adding more languages improves unsupervised multilingual part-of-speech tagging: A Bayesian non-parametric approach. In Proc. of NAACL. find
  55. K. Toutanova, D. Klein, C. D. Manning, and Y. Singer. 2003. Feature-rich part-of-speech tagging with a cyclic dependency network. In Proc. of HLT-NAACL. find our card
  56. L. Van der Beek, G. Bouma, R. Malouf, and G. Van Noord. 2002. The Alpino dependency treebank. Language and Computers, 45(1):8–22. find
  57. C. Xi and R. Hwa. 2005. A backoff model for bootstrapping resources for non-English languages. In Proc. of HLT-EMNLP. find
  58. D. Yarowsky and G. Ngai. 2001. Inducing multilingual POS taggers and NP bracketers via robust projection across aligned corpora. In Proc. of NAACL. find
  59. D. Zeman and P. Resnik. 2008. Cross-language parser adaptation between related languages. In NLP for Less Privileged Languages. find our card
  60. D. Zeman. 2008. Reusable tagset conversion using tagset drivers. Proc. of LREC. find our card
  61. Fg, Fh, Fi, Fp, Fa, Fc, Fd, Fe Fs, Fx, Fz, fa fc, fg, fp Punct PU . G-BA, FW, IJ, LB, ON, SB, X X –, I-, Y-, I, i, find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/