Skip to main content

Extending Feature Decay Algorithms Using Alignment Entropy

Part of the Lecture Notes in Computer Science book series (LNAI,volume 10341)

Abstract

In machine-learning applications, data selection is of crucial importance if good runtime performance is to be achieved. Feature Decay Algorithms (FDA) have demonstrated excellent performance in a number of tasks. While the decay function is at the heart of the success of FDA, its parameters are initialised with the same weights. In this paper, we investigate the effect on Machine Translation of assigning more appropriate weights to words using word-alignment entropy. In experiments on German to English, we show the effect of calculating these weights using two popular alignment methods, GIZA++ and FastAlign, using both automatic and human evaluations. We demonstrate that our novel FDA model is a promising research direction.

Keywords

  • Data selection
  • Machine translation
  • Mathematical foundations

This is a preview of subscription content, access via your institution.

Buying options

Chapter
USD   29.95
Price excludes VAT (USA)
  • DOI: 10.1007/978-3-319-69365-1_14
  • Chapter length: 13 pages
  • Instant PDF download
  • Readable on all devices
  • Own it forever
  • Exclusive offer for individuals only
  • Tax calculation will be finalised during checkout
eBook
USD   44.99
Price excludes VAT (USA)
  • ISBN: 978-3-319-69365-1
  • Instant PDF download
  • Readable on all devices
  • Own it forever
  • Exclusive offer for individuals only
  • Tax calculation will be finalised during checkout
Softcover Book
USD   59.99
Price excludes VAT (USA)
Fig. 1.

Notes

  1. 1.

    Note that we were unable to calculate statistical significance for the CHRF metric. Note too that prior to tuning, statistically significant improvements were seen for GIZA++ over FastAlign for BLEU, NIST, TER and METEOR.

  2. 2.

    In experiments before tuning (excluded here for reasons of space), the METEOR and CHRF scores of the output of the system executed with GIZA++ did outperform the baseline system before tuning.

  3. 3.

    Why is it the case that better scores are more likely with the METEOR evaluation metric? This measure evaluates a hypothesis against a reference calculating sentence-level similarity scores. In so doing it searches for all the possible matches of the words between the two sentences. The words can match (i) if they are the same, (ii) if they have the same stem, (iii) if they are synonyms, or (iv) if they are found as a match in a paraphrase table. Therefore, this metric takes into consideration not only the n-grams, but also the semantic of the words. As the human evaluation shows, many semantically equivalent translations are output by our GIZA++ system, which are penalised by most automatic metrics, but not by METEOR.

References

  1. Koehn, P., Och, F., Marcu, D.: Statistical phrase-based translation. In: Proceedings of Conference combining Human Language Technology Conference series and the North American Chapter of the Association for Computational Linguistics Conference serie. (edmonton, Canada), pp. 48–54 (2003)

    Google Scholar 

  2. Papineni, K., Roukos, S., Ward, T., Zhu, W.-J.: Bleu: a method for automatic evaluation of machine translation. In: Proceedings of the 40th ACL, Philadelphia, PA, USA, pp. 311–318 (2002)

    Google Scholar 

  3. Koehn, P.: Europarl: a parallel corpus for statistical machine translation. In: MT Summit X Conference Proceedings: The Tenth Machine Translation Summit, Phuket, Thailand, pp. 79–86 (2005)

    Google Scholar 

  4. Ozdowska, S., Way, A.: Optimal bilingual data for French-English PB-SMT. In: Proceedings of the 13th Annual Meeting of the European Association for Machine Translation, Barcelona, Spain, pp. 96–103 (2009)

    Google Scholar 

  5. Biçici, E., Yuret, D.: Instance selection for machine translation using feature decay algorithms. In: Proceedings of the Sixth Workshop on Statistical Machine Translatio, Edinburgh, Scotland, pp. 272–283 (2011)

    Google Scholar 

  6. Dyer, C., Chahuneau, V., Smith, N.: A simple, fast, and effective reparameterization of IBM model 2. In: Proceedings of Human Language Technologies: Conference of the North American Chapter of the Association of Computational Linguistics, Proceedings, Atlanta, Georgia, USA, pp. 644–648 (2013)

    Google Scholar 

  7. Och, F., Ney, H.: A systematic comparison of various statistical alignment models. Comput. Linguist. 29(1), 19–51 (2003)

    CrossRef  MATH  Google Scholar 

  8. Shterionov, D., Du, J., Palminteri, M.A., Casanellas, L., ODowd, T., Way, A.: Improving KantanMT training efficiency with FastAlign. In: Proceedings of AMTA 2016, Austin, TX, pp. 222–231 (2016)

    Google Scholar 

  9. Eetemadi, S., Lewis, W., Toutanova, K., Radha, H.: Survey of data-selection methods in statistical machine translation. Mach. Transl. 29(3–4), 189–223 (2015)

    CrossRef  Google Scholar 

  10. Biçici, E., Liu, Q., Way, A.: ParFDA for fast deployment of accurate statistical machine translation systems, benchmarks, and statistics. In: Proceedings of the Tenth Workshop on Statistical Machine Translation, Lisbon, Portugal, pp. 74–78 (2015)

    Google Scholar 

  11. Brown, P.F., Della Pietra, S.A., Della Pietra, V.J., Mercer, R.L.: The mathematics of statistical machine translation: parameter estimation. Comput. Linguist. 19, 263–311 (1993)

    Google Scholar 

  12. Koehn, P., Hoang, H., Birch, A., Callison-Burch, C., Federico, M., Bertoldi, N., Cowan, B., Shen, W., Moran, C., Zens, R., Dyer, C., Bojar, O., Constantin, A., Herbst, E.: Moses: open source toolkit for SMT. In: Proceedings of 45th Annual Meeting of the ACL on Interactive Poster & Demonstration Sessions, (Prague, Czech Republic), pp. 177–180 (2007)

    Google Scholar 

  13. Och, F.: Minimum error rate training in statistical machine translation. In: ACL-2003: 41st Annual Meeting of the Association for Computational Linguistics, Proceedings, Sapporo, Japan, pp. 160–167 (2003)

    Google Scholar 

  14. Clark, J.H., Dyer, C., Lavie, A., Smith, N.A.: Better hypothesis testing for statistical machine translation: controlling for optimizer instability. In: Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, vol. 2, Short Papers, Portland, Oregon, pp. 176–181 (2011)

    Google Scholar 

  15. Heafield, K.: KenLM: faster and smaller language model queries. In: Proceedings of the Sixth Workshop on Statistical Machine Translation, Edinburgh, Scotland, pp. 187–197 (2011)

    Google Scholar 

  16. Doddington, G.: Automatic evaluation of machine translation quality using n-gram co-occurrence statistics. In: Proceedings of the Second International Conference on Human Language Technology Research, San Diego, CA, pp. 138–145 (2002)

    Google Scholar 

  17. Snover, M., Dorr, B., Schwartz, R., Micciulla, L., Makhoul, J.: A study of translation edit rate with targeted human annotation. In: Proceedings of the 7th Conference of the Association for Machine Translation in the Americas, Cambridge, Massachusetts, USA, pp. 223–231 (2006)

    Google Scholar 

  18. Banerjee, S., Lavie, A.: Meteor: an automatic metric for MT evaluation with improved correlation with human judgments. In: Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization, Ann Arbor, Michigan, pp. 65–72 (2005)

    Google Scholar 

  19. Popovic, M.: chrF: character n-gram F-score for automatic MT evaluation. In: Proceedings of the Tenth Workshop on Statistical Machine Translation, Lisbon, Portugal, pp. 392–395 (2015)

    Google Scholar 

  20. Koehn, P.: Statistical significance tests for machine translation evaluation. In: Proceedings of EMNLP 2004, Barcelona, Spain, pp. 388–395 (2004)

    Google Scholar 

  21. He, Y., Way, A.: Improving the objective function in minimum error rate training. In: Proceedings of the Twelfth Machine Translation Summit, Ottawa, Canada, pp. 238–245 (2009)

    Google Scholar 

Download references

Acknowledgements

This research is supported by the ADAPT Centre for Digital Content Technology, funded under the SFI Research Centres Programme (Grant 13/RC/2106) and co-funded under the European Regional Development Fund, and the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement PIAP-GA-2012-324414 (AbuMaTran).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Alberto Poncelas .

Editor information

Editors and Affiliations

Rights and permissions

Reprints and Permissions

Copyright information

© 2017 Springer International Publishing AG

About this paper

Verify currency and authenticity via CrossMark

Cite this paper

Poncelas, A., Way, A., Toral, A. (2017). Extending Feature Decay Algorithms Using Alignment Entropy. In: Quesada, J., Martín Mateos , FJ., López Soto, T. (eds) Future and Emerging Trends in Language Technology. Machine Learning and Big Data. FETLT 2016. Lecture Notes in Computer Science(), vol 10341. Springer, Cham. https://doi.org/10.1007/978-3-319-69365-1_14

Download citation

  • DOI: https://doi.org/10.1007/978-3-319-69365-1_14

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-319-69364-4

  • Online ISBN: 978-3-319-69365-1

  • eBook Packages: Computer ScienceComputer Science (R0)