Real-Time Statistical Speech Translation

Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 275)


This research investigates the Statistical Machine Translation approaches to translate speech in real time automatically. Such systems can be used in a pipeline with speech recognition and synthesis software in order to produce a real-time voice communication system between foreigners. We obtained three main data sets from spoken proceedings that represent three different types of human speech. TED, Europarl, and OPUS parallel text corpora were used as the basis for training of language models, for developmental tuning and testing of the translation system. We also conducted experiments involving part of speech tagging, compound splitting, linear language model interpolation, TrueCasing and morphosyntactic analysis. We evaluated the effects of variety of data preparations on the translation results using the BLEU, NIST, METEOR and TER metrics and tried to give answer which metric is most suitable for PL-EN language pair.


Machine translation Speech translation Machine learning NLP Knowledge-free learning 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Koehn, P., Hoang, H.: Moses: Open Source Toolkit for Statistical Machine Translation, Prague (2007)Google Scholar
  2. 2.
    Marasek, K.: TED Polish-to-English translation system for the IWSLT 2012. In: IWSLT 2012, Hong Kong (2012)Google Scholar
  3. 3.
    Costa-Jussa, M., Fonollosa, J.: Using linear interpolation and weighted reordering hypotheses in the Moses system, Barcelona, Spain (2010)Google Scholar
  4. 4.
    Stolcke, A.: SRILM – An Extensible Language Modeling Toolkit. In: INTERSPEECH (2002)Google Scholar
  5. 5.
    Hsu, P., Glass, J.: Iterative Language Model Estimation: Efficient Data Structure & Algorithms, Cambridge, USA (2008)Google Scholar
  6. 6.
    Bojar, O.: Rich Morphology and What Can We Expect from Hybrid Approaches to MT. In: LIHMT 2011 (2011)Google Scholar
  7. 7.
    Radziszewski, A.: A tiered CRF tagger for Polish. In: Bembenik, R., Skonieczny, Ł., Rybiński, H., Kryszkiewicz, M., Niezgódka, M. (eds.) Intell. Tools for Building a Scientific Information. SCI, vol. 467, pp. 215–230. Springer, Heidelberg (2013)Google Scholar
  8. 8.
    Koehn, P., Hoang, H.: Factored Translation Models, Scotland, United Kingdom (2007)Google Scholar
  9. 9.
    Ratnaparkhi, A.: A Maximum Entropy Part-Of-Speech Tagger, Pennsylvania (1996)Google Scholar
  10. 10.
    Holz, F., Biemann, C.: Unsupervised and knowledge-free learning of compound splits and periphrases. In: Gelbukh, A. (ed.) CICLing 2008. LNCS, vol. 4919, pp. 117–127. Springer, Heidelberg (2008)Google Scholar
  11. 11.
    Cer, D., Manning, C., Jurafsky, D.: The Best Lexical Metric for Phrase-Based Statistical MT System Optimization. Stanford, USA (2010)Google Scholar
  12. 12.
    Gao, Q., Vogel, S.: Parallel Implementations of Word Alignment Tool (2008)Google Scholar
  13. 13.
    Heafield, K.: KenLM: Faster and smaller language model queries. Association for Computational Linguistics (2011)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Department of MultimediaPolish Japanese Institute of Information TechnologyWarsawPoland

Personalised recommendations