Elamur Research⌕
☰
home / papers / 2017 / Personalized Machine Translation: Preserving Original Author Traits

Personalized Machine Translation: Preserving Original Author Traits

Paper · 2017 · Ella Rabinovich, Raj Nath Patel, Shachar Mirkin et al.

Summary

The paper examines whether an author's demographic traits, specifically gender, survive translation, and whether machine translation can be adjusted to keep them. The authors build a new resource: the en-fr and en-de Europarl parallel corpora, with each speaker (a Member of the European Parliament) labelled for gender and for age at the time of the session. They also reuse a small TED talks dataset from Mirkin et al. (2015). Gender classifiers are trained and tested on original texts, on human translations and on SMT output, for both language pairs in both directions. Gender is easy to detect in originals but becomes much weaker after translation. Surprisingly, it is weaker after human translation than after SMT. The authors then build "gender-aware" SMT systems using simple domain-adaptation techniques. These systems keep more gender markers without lowering general translation quality. The work matters because it frames personalization of MT as a domain problem. It also shows that translations reflect the source language's gender signals more than the target language's.

Key points

  • Resource: The annotated Europarl data is restricted to sentences originally spoken in English, French or German. Sizes are about 144K (en-fr), 107K (fr-en), 162K (en-de) and 131K (de-en) sentence-pairs. A list of 3,586 MEPs with metadata is also released at cl.haifa.ac.il/projects/pmt.
  • Gender labelling: Three sources were combined:
    • Wikidata covered 2,618 MEPs (73%).
    • Genderize, using first name plus represented country, kept 2,730 MEPs after removing 55 low-confidence cases.
    • AlchemyVision, applied to MEP photos, gave labels for 2,236 MEPs, with predictions below 0.9 confidence discarded.
  • Classification setup: The design accounts for the finding that the translation process leaves a stronger signal than the author's gender.
  • Feature analysis: Discriminative function words were selected with InfoGain. Example markers include also, very and perhaps (English Europarl), je, vous and dire (French), and wir, man and für (German). In TED, the markers are mostly pronouns.
  • Source traits carry over: In fr-en translations, English markers mirror the French ones. For example, must becomes a female marker because it renders doit, even though must signals male speech in original English.

Interesting ideas

  • I1. Personalization as domain adaptation: Treating each author group (here, gender) as a separate "domain" lets existing adaptation tools serve personalization. The same trick could apply to dialect, register or speaker age.
  • I2. Cross-checking cheap metadata sources: A knowledge base, a name-based predictor and image recognition were combined, each filtered by a confidence threshold. This is a practical recipe for enriching any speaker-attributed corpus with demographic labels.
  • I3. Using a trait classifier as an evaluation metric: How well a trait classifier works on translated output measures what a translation preserves. This complements BLEU-style quality scores with a check on style faithfulness.
  • I4. Inspecting the sentences that flip: Because all systems translate exactly the same sentences, the authors examine those the baseline misclassifies but the personalized model gets right. One example is vraiment rendered as "exactly" instead of "really." This is a clean way to see concretely what an adaptation changes.
  • I5. Tracking where markers come from: Comparing InfoGain marker weights across originals and translations shows whether a feature comes from the source or the target language. This is useful for any study of translationese or of data built through cross-lingual transfer.

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

 Institutions

🎓University of SheffieldSheffield · United Kingdom76 works · 20,808 citations🎓University of HaifaHaifa · Israel27 works · 871 citations🔬IBM Research - HaifaHaifa · Israel5 works · 1,112 citations

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 works

🧭 Topics 1Research subjects and languages the work is about.Machine translation · 958 works

Cite

Cite as Ella Rabinovich, Raj Nath Patel, Shachar Mirkin, Lucia Specia and Shuly Wintner. 2017. Personalized Machine Translation: Preserving Original Author Traits. Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers.

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

@inproceedings{rabinovich2017personalized,
  title = {{Personalized Machine Translation: Preserving Original Author Traits}},
  author = {Ella Rabinovich, Raj Nath Patel, Shachar Mirkin, Lucia Specia and Shuly Wintner},
  booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers},
  year = {2017},
  doi = {10.18653/v1/e17-1101},
  url = {https://doi.org/10.18653/v1/e17-1101}
}

References

40 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Roee Aharoni, Moshe Koppel, and Yoav Goldberg. 2014. Automatic detection of machine translated text and translation quality estimation. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, ACL 2014, June 22-27, 2014, Baltimore, MD, USA, Volume 2: Short Papers,… find
  2. Yuki Arase and Ming Zhou. 2013. Machine translation detection from monolingual web-text. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, ACL 2013, 4-9 August 2013, Sofia, Bulgaria, Volume 1: Long Papers, pages 1597–1607. find
  3. Marco Baroni and Silvia Bernardini. 2006. A new approach to the study of Translationese: Machine-learning the difference between original and translated text. Literary and Linguistic Computing, 21(3):259–274, September. find
  4. John D. Burger, John Henderson, George Kim, and Guido Zarrella. 2011. Discriminating gender on twitter. In Proceedings of the Conference on Empirical Methods in Natural Language Processing, EMNLP ’11, pages 1301–1309, Stroudsburg, PA, USA. Association for Computational Linguistics.
  5. Mauro Cettolo, Christian Girardi, and Marcello Federico. 2012. Wit 3 : Web inventory of transcribed and translated talks. In Proceedings of the 16 th Conference of the European Association for Machine Translation (EAMT), pages 261–268, Trento, Italy, May. find
  6. Stanley F. Chen and Joshua Goodman. 1996. An empirical study of smoothing techniques for language modeling. In Proceedings of ACL.
  7. Costanza Cucchi. 2012. Dialogic features in EU non-native parliamentary debates. Review of the Air Force Academy, 11(3):5–14. find
  8. Lei Cui, Dongdong Zhang, Shujie Liu, Mu Li, and Ming Zhou. 2013. Bilingual data cleaning for smt using graph-based random walk. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, pages 340–345, Sofia, Bulgaria. find
  9. Hoang Cuong and Khalil Sima’an. 2015. Latent domain word alignment for heterogeneous corpora. In Proceedings of the Annual Meeting of the North American Chapter of the Association for Computational Linguistics, pages 398–408, Denver, USA. find
  10. Richard S Forsyth and Phoenix WY Lam. 2014. Found in translation: To what extent is authorial discriminability preserved by translators? Literary and Linguistic Computing, 29(2):199–217. find
  11. Martin Gellerstam. 1986. Translationese in Swedish novels translated from English. In Lars Wollin and Hans Lindquist, editors, Translation Studies in Scandinavia, pages 88–95. CWK Gleerup, Lund. find
  12. Robert M Gray. 1990. Entropy and information. In Entropy and Information Theory, pages 21–55. Springer. find
  13. Mark Hall, Eibe Frank, Geoffrey Holmes, Bernhard Pfahringer, Peter Reutemann, and Ian H. Witten. 2009. The WEKA data mining software: an update. SIGKDD Explorations, 11(1):10–18. find
  14. Eva Hasler, Barry Haddow, and Philipp Koehn. 2014. Dynamic topic adaptation for smt using distributional profiles. In Proceedings of the Ninth Workshop on Statistical Machine Translation, pages 445– 456, Baltimore, USA. find
  15. Kenneth Heafield. 2011. KenLM: faster and smaller language model queries. In Proceedings of the EMNLP 2011 Sixth Workshop on Statistical Machine Translation, pages 187–197, Edinburgh, Scotland, United Kingdom, July.
  16. Dirk Hovy. 2015. Demographic factors improve classification performance. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 752–762, Beijing, China.
  17. Anders Johannsen, Dirk Hovy, and Anders Søgaard. 2015. Cross-lingual syntactic variation over age and gender. In Proceedings of the Nineteenth Conference on Computational Natural Language Learning, pages 103–112, Beijing, China, July. Association for Computational Linguistics. find
  18. Ian Jolliffe. 2002. Principal component analysis. Wiley Online Library. find
  19. S.S. Keerthi, S.K. Shevade, C. Bhattacharyya, and K.R.K. Murthy. 2001. Improvements to platt’s smo algorithm for svm classifier design. Neural Computation, 13(3):637–649. find
  20. Philipp Koehn, Hieu Hoang, Alexandra Birch, Chris Callison-Burch, Marcello Federico, Nicola Bertoldi, Brooke Cowan, Wade Shen, Christine Moran, Richard Zens, Chris Dyer, Ondřej Bojar, Alexandra Constantin, and Evan Herbst. 2007. Moses: Open source toolkit for statistical machine translation. In Pro… find
  21. Gerard Lynch. 2014. A supervised learning approach towards profiling the preservation of authorial style in literary translations. In COLING, pages 376–386. find
  22. Christopher D. Manning, Mihai Surdeanu, John Bauer, Jenny Finkel, Steven J. Bethard, and David Mc-Closky. 2014. The Stanford CoreNLP natural language processing toolkit. In Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 55–60, Baltim…
  23. Shachar Mirkin and Jean-Luc Meunier. 2015. Personalized machine translation: Predicting translational preferences. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 2019–2025, Lisbon, Portugal, September. Association for Computational Linguistics. find
  24. Shachar Mirkin, Scott Nowson, Caroline Brun, and Julien Perez. 2015. Motivating personality-aware machine translation. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1102–1108, Lisbon, Portugal, September. Association for Computational Linguistics. find
  25. Matthew L. Newman, Carla J. Groom, Lori D. Handelman, and James W. Pennebaker. 2008. Gender differences in language use: An analysis of 14,000 text samples. Discourse Processes, 45(3):211–236. find
  26. Franz Josef Och. 2003. Minimum error rate training in statistical machine translation. In Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - Volume 1 (ACL 2003), ACL ’03, pages 160–167. find
  27. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: a Method for Automatic Evaluation of Machine Translation. In Proceedings of ACL, Philadelphia, Pennsylvania, USA.
  28. Ella Rabinovich and Shuly Wintner. 2015. Unsupervised identification of translationese. Transactions of the Association for Computational Linguistics, 3:419–432. find
  29. Ella Rabinovich, Shuly Wintner, and Ofek Luis Lewinsohn. 2015. The haifa corpus of translationese. arXiv preprint arXiv:1509.03611. — [arXiv:1509.03611](https://arxiv.org/abs/1509.03611) find
  30. Philipp Koehn. 2005. Europarl: A parallel corpus for statistical machine translation. MT Summit. find
  31. Jonathan Schler, Moshe Koppel, Shlomo Argamon, and James W. Pennebaker. 2006. Effects of age and gender on blogging. In Computational Approaches to Analyzing Weblogs, Papers from the 2006 AAAI Spring Symposium, Technical Report SS-06-03, Stanford, California, USA, March 27-29, 2006, pages 199–205. find
  32. Moshe Koppel, Shlomo Argamon, and Anat Rachel Shimoni. 2002. Automatically categorizing written texts by author gender. LLC, 17(4):401–412. find
  33. Yanir Seroussi, Ingrid Zukerman, and Fabian Bohnert. 2014. Authorship attribution with topic models. Comput. Linguist., 40(2):269–310. find
  34. Miriam Shlesinger, Moshe Koppel, Noam Ordan, and Brenda Malkiel. 2009. Markers of translator gender: do they really matter? Internet. Disponı́vel em http://u.cs. biu. ac. il/˜ koppel/Publications. ht ml (consultado em 31de março de 2011). ↗
  35. Michel Simard. 2014. Clean data for training statistical mt: the case of mt contamination. In Proceedings of the Eleventh Conference of the Association for Machine Translation in the Americas, pages 69– 82, Vancouver, BC, Canada. find
  36. Sherry Simon. 2003. Gender in translation. Routledge. find
  37. Vered Volansky, Noam Ordan, and Shuly Wintner. 2015. On the features of translationese. Digital Scholarship in the Humanities, 30(1):98–118, April. find
  38. Svitlana Volkova, Theresa Wilson, and David Yarowsky. 2013. Exploring demographic language variations to improve multilingual sentiment analysis in social media. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1815–1827, Seattle, Washington, USA.
  39. Luise Von Flotow. 2010. Gender in translation. Handbook of Translation Studies, 1:129–133. find
  40. Denny Vrandečić and Markus Krötzsch. 2014. Wikidata: A free collaborative knowledgebase. Commun. ACM, 57(10):78–85, September. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/rabinovich-2017-personalized-machine-translation-preserving-original/