Skip to main content
Springer Nature Link
Log in

Automatic translation memory cleaning

  • Published:
Machine Translation

    We’re sorry, something doesn't seem to be working properly.

    Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.

Abstract

We address the problem of automatically cleaning a translation memory (TM) by identifying problematic translation units (TUs). In this context, we treat as “problematic TUs” those containing useless translations from the point of view of the user of a computer-assisted translation tool. We approach TM cleaning both as a supervised and as an unsupervised learning problem. In both cases, we take advantage of Translation Memory open-source purifier, an open-source TM cleaning tool also presented in this paper. The two learning paradigms are evaluated on different benchmarks extracted from MyMemory, the world’s largest public TM. Our results indicate the effectiveness of the supervised approach in the ideal condition in which labelled training data is available, and the viability of the unsupervised solution for challenging situations in which training data is not accessible.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2

Similar content being viewed by others

Explore related subjects

Discover the latest articles, news and stories from top researchers in related subjects.

Notes

  1. In addition to these components, CAT tools can be equipped with concordancers, terminology databases, spell/grammar checkers, indexers and project management functions.

  2. http://rgcl.wlv.ac.uk/nlp4tm2016/shared-task/.

  3. The translation in (d) contains “somministARzione” instead of “somministRAzione”.

  4. It is worth remarking that not all existing TMs are private resources carefully constructed by expert human translators. Some of them are collaboratively built by anonymous contributors and can also include TUs automatically extracted from the Web. In such cases, major translation errors are quite frequent.

  5. http://www.xbench.net/.

  6. https://mymemory.translated.net/.

  7. https://ec.europa.eu/jrc/en/language-technologies/dgt-translation-memory.

  8. Although not usable in practice, however, the simplest “majority voting” baseline marking all the TUs as “good” achieves even better results on the same highly imbalanced data.

  9. http://alt.qcri.org/semeval2016/task1/.

  10. https://github.com/hlt-mt/TMOP.

  11. For instance, judging the usefulness of a TU whose target side has missing/extra words is a highly subjective task. It is likely that the perceived severity of these errors will be inversely proportional to sentence length.

  12. The evaluation on the third subtask, which corresponds to a multi-class interpretation of the problem, is not discussed for the sake of conciseness.

  13. In the NLP4TM shared task data, the threshold is set to 2.5.

  14. A detailed summary of the FBK HLT-MT submissions to the 1st translation memory cleaning shared task is available at http://rgcl.wlv.ac.uk/wp-content/uploads/2016/05/fbkhltmt-workingnote.

  15. As stated in Sect. 1, we call this method “unsupervised” to convey the idea that, although it relies on supervised learning algorithms, it bypasses the need for the supervision provided by manual labels.

  16. Indeed, each model is obtained with different features (e.g. the A group) and by learning from training data having different label distributions (e.g. inferred using B and C).

  17. https://www.bing.com/translator/.

  18. All the improvements over the MT-based system are statistically significant (\(\rho <0.05\) measured by approximate randomization), while only the result for Z \(=\) 50 K and k \(=\) 15 K is statistically significantly better than the Barbu15 classifier.

  19. In total, this corresponds to 30% of the whole test set.

References

  • Abdul Rauf S, Schwenk H (2011) Parallel sentence generation from comparable corpora for improved SMT. Mach Transl 25(4):341–375

    Article  Google Scholar 

  • Arthern P (1979) Machine translation and computerized terminology systems: a translator’s viewpoint. In: Translating and the computer, proceedings of a seminar, London, UK, pp 77–108

  • Barbu E (2015) Spotting false translation segments in translation memories. In: Proceedings of the workshop on natural language processing for translation memories, Hissar, Bulgaria, pp 9–16

  • Barbu E, Parra Escartín C, Bentivogli L, Negri M, Turchi M, Federico M, Mastrostefano L, Orasan C (2016) 1st shared task on automatic translation memory cleaning. In: Proceedings of the 2nd Workshop on natural language processing for translation memories (NLP4TM 2016). Portorož, Slovenia, pp 1–5

  • Biçici E, Dymetman M (2008) Dynamic translation memory: using statistical machine translation to improve translation memory fuzzy matches. In: Proceedings of the 9th international conference on computational linguistics and intelligent text processing, CICLing’08, Haifa, Israel, pp 454–465

  • Bloodgood M, Strauss B (2014) Translation memory retrieval methods. In: Proceedings of the 14th conference of the European chapter of the association for computational linguistics, Gothenburg, Sweden, pp 202–210

  • Brown PF, Della Pietra SA, Della Pietra VJ, Mercer RL (2003) The mathematics of statistical machine translation: parameter estimation. Comput Linguist 19(2):263–311

    Google Scholar 

  • Burchardt A, Lommel A (2014) Practical guidelines for the use of MQM in scientific research on translation quality. Technical report, DFKI, Berlin, Germany

  • Camargo de Souza JG, Buck C, Turchi M, Negri M (2013) FBK-UEdin participation to the WMT13 quality estimation shared task. In: Proceedings of the eighth workshop on statistical machine translation, Sofia, Bulgaria, pp 352–358

  • Chatzitheodoroou K (2015) Improving translation memory fuzzy matching by paraphrasing. In: Proceedings of the workshop on natural language processing for translation memories, Hissar, Bulgaria, pp 24–30

  • Chu C, Nakazawa T, Kurohashi S (2013) Chinese–Japanese parallel sentence extraction from quasi–comparable corpora. In: Proceedings of the sixth workshop on building and using comparable corpora, Sofia, Bulgaria, pp 34–42

  • Cotterell R, Schütze H, Eisner J (2016) Morphological smoothing and extrapolation of word embeddings. In: Proceedings of the 54th annual meeting of the association for computational linguistics (volume 1: long papers), Berlin, Germany, pp 1651–1660

  • Denkowski M, Hanneman G, Lavie A (2012) The CMU-avenue French–English translation system. In: Proceedings of the seventh workshop on statistical machine translation, Montréal, Canada, pp 261–266

  • Dyer C, Clark J, Lavie A, Smith NA (2011) Unsupervised word alignment with arbitrary features. In: Proceedings of the 49th annual meeting of the association for computational linguistics: human language technologies-volume 1, Portland, Oregon, USA, pp 409–419

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

    Article  Google Scholar 

  • Gao Q, Vogel S (2008) Parallel implementations of word alignment tool. In: Proceedings of the ACL 2008 software engineering, testing, and quality assurance workshop, Columbus, Ohio, USA, pp 49–57

  • Geurts P, Ernst D, Wehenkel L (2006) Extremely randomized trees. Mach Learn 63(1):3–42

    Article  MATH  Google Scholar 

  • Gupta R, Bechara H, Orasan C (2014) Intelligent translation memory matching and retrieval metric exploiting linguistic technology. In: Proceedings of translating and the computer 36, London, UK, pp 86–89

  • Gupta R, Orasan C, Zampieri M, Vela M, Van Genabith J (2015) Can translation memories afford not to use paraphrasing? In: Proceedings of the 18th annual conference of the European association for machine translation, Antalya, Turkey, pp 35–42

  • Khadivi S, Ney H (2005) Automatic filtering of bilingual corpora for statistical machine translation. In: Proceedings of natural language processing and information systems, 10th international conference on applications of natural language to information systems, Alicante, Spain, pp 263–274

  • Koehn P, Senellart J (2010) Convergence of translation memory and statistical machine translation. In: Proceedings of AMTA workshop on MT research and the translation industry, Denver, CO, USA, pp 21–31

  • Levenshtein VI (1966) Binary codes capable of correcting deletions, insertions, and reversals. Sov Phys Dokl 10:707–710

    MathSciNet  MATH  Google Scholar 

  • Lommel A (2015) Multidimensional quality metrics (MQM) definition. Technical report, DFKI, Berlin, Germany

  • Lui M, Baldwin T (2012) langid.py: an off-the-shelf language identification tool. In: Proceedings of the ACL 2012 system demonstrations, Jeju Island, Korea, pp 25–30

  • Ma Y, He Y, Way A, Van Genabith J (2011) Consistent translation using discriminative learning: a translation memory-inspired approach. In: Proceedings of the 49th annual meeting of the association for computational linguistics: human language technologies, volume 1, Portland, Oregon, USA, pp 1239–1248

  • Meinshausen N, Bühlmann P (2010) Stability selection. J R Stat Soc B Stat Methodol 72(4):417–473

    Article  MathSciNet  Google Scholar 

  • Munteanu DS, Marcu D (2005) Improving machine translation performance by exploiting non-parallel corpora. Comput Linguist 31(4):477–504

    Article  Google Scholar 

  • Nakazawa T, Kurohashi S (2011) Bayesian subtree alignment model based on dependency trees. In: Proceedings of 5th international joint conference on natural language processing, Chiang Mai, Thailand, pp 794–802

  • Negri M, Marchetti A, Mehdad Y, Bentivogli L, Giampiccolo D (2012) Semeval-2012 task 8: cross-lingual textual entailment for content synchronization. In: Proceedings of the 6th international workshop on semantic evaluation (SemEval 2012), Montréal, Canada, pp 399–407

  • Noreen EW (1989) Computer intensive methods for testing hypothesis. An introduction. Wiley, New York

    Google Scholar 

  • Rarrick S, Quirk C, Lewis W (2011) MT detection in web-scraped parallel corpora. In: MT summit XIII: the thirteenth machine translation summit, Xiamen, China, pp 422–429

  • Riesa J, Marcu D (2012) Automatic parallel fragment extraction from noisy data. In: Proceedings of the 2012 conference of the North American chapter of the association for computational linguistics: human language technologies, Montréal, Canada, pp 538–542

  • Sikes R (2007) Fuzzy matching in theory and practice. Multilingual 18(6):39–43

    Google Scholar 

  • Snover M, Dorr B, Schwartz R, Micciulla L, Makhoul J (2006) A study of translation edit rate with targeted human annotation. In: AMTA 2006: proceedings of the 7th conference of the association for machine translation in the Americas, visions for the future of machine translation, Cambridge, Massachusetts, USA, pp 223–231

  • Søgaard A, Agić V, Martínez Alonso H, Plank B, Bohnet B, Johannsen A (2015) Inverted indexing for cross-lingual NLP. In: Proceedings of the 53rd annual meeting of the association for computational linguistics and the 7th international joint conference on natural language processing (volume 1: long papers), Beijing, China, pp 1713–1722

  • Specia L, Cancedda N, Dymetman M, Turchi M, Cristianini N (2009) Estimating the sentence-level quality of machine translation systems. In: Proceedings of the 13th annual conference of the European association for machine translation (EAMT-2009), Barcelona, Spain, pp 28–35

  • Tillmann C (2009) A beam-search extraction algorithm for comparable data. In: Proceedings of the ACL-IJCNLP 2009 conference short papers, Singapore, pp 225–228

  • Turchi M, Negri M, Federico M (2013) Coping with the subjectivity of human judgements in MT quality estimation. In: Proceedings of the eighth workshop on statistical machine translation, Sofia, Bulgaria, pp 240–251

  • Turchi M, Negri M, Federico M (2014) Data-driven annotation of binary MT quality estimation corpora based on human post-editions. Mach Transl 28(3):281–308

    Article  Google Scholar 

  • Vanallemeersch T, Vandeghinste V (2014) Improving fuzzy matching through syntactic knowledge. In: Proceedings of translating and the computer 36, London, pp 217–227

  • Vanallemeersch T, Vandeghinste V (2015) Assessing linguistically aware fuzzy matching in translation memories. In: Proceedings of the 18th annual conference of the European association for machine translation, Antalya, Turkey, pp 153–160

  • Wang K, Zong C, Su KY (2013) Integrating translation memory into phrase-based machine translation during decoding. In: Proceedings of the 51st annual meeting of the association for computational linguistics (volume 1: long papers), Sofia, Bulgaria, pp 11–21

  • Yeh A (2000) More accurate tests for the statistical significance of result differences. In: The 18th international conference on computational linguistics, COLING 2000 in Europe, proceedings of the conference, volume 2, Saarbrücken, Germany, pp 947–953

  • Zhechev V, Van Genabith J (2010) Seeding statistical machine translation with translation memory output through tree-based structural alignment. In: Proceedings of the 4th workshop on syntax and structure in statistical translation, Beijing, China, pp 43–51

Download references

Acknowledgements

This work has been partially supported by the EC-funded project ModernMT (H2020 Grant Agreement No. 645487). The work carried out at FBK by Masoud Jalili Sabet was sponsored by the European Association for Machine Translation, through the EAMT summer internships 2015 program. The authors would also like to thank Translated for providing a dump of MyMemory, Eduard Barbu for running his system on our data, and the anonymous reviewers for their insightful comments.

Author information

Authors and Affiliations

  1. Fondazione Bruno Kessler, Povo, Trento, Italy

    Matteo Negri, Marco Turchi & Marcello Federico

  2. Fondazione Bruno Kessler, Università degli Studi di Trento, Povo, Trento, Italy

    Duygu Ataman

  3. School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran

    Masoud Jalili Sabet

Authors
  1. Matteo Negri
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Duygu Ataman
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Masoud Jalili Sabet
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Marco Turchi
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Marcello Federico
    View author publications

    You can also search for this author inPubMed Google Scholar

Corresponding author

Correspondence to Matteo Negri.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Negri, M., Ataman, D., Sabet, M.J. et al. Automatic translation memory cleaning. Machine Translation 31, 93–115 (2017). https://doi.org/10.1007/s10590-017-9191-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10590-017-9191-5

Keywords