Transducer-learning experiments on language understanding

  • David Picó
  • Enrique Vidal
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1433)


The interest in using Finite-State Models in a large variety of applications is recently growing as more powerful techniques for learning them from examples have been developed. Language Understanding can be approached this way as a problem of language translation in which the target language is a formal language rather than a natural one. Finite-state transducers are used to model the translation process, and are automatically learned from training data consisting of pairs of natural-language/formal-language sentences. The need for training data is dramatically reduced by performing a two-level learning process based on lexical/phrase categorization. Successful experiments are presented on a task consisting in the “understanding” of Spanish natural-language sentences describing dates and times, where the target formal language is the one used in the popular Unix command “at”.


Machine Translation Language Translation Training Corpus Training Pair Input Sentence 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    J.C Amengual, J.B. Benedí, F. Casacuberta, A. Castaño, A. Castellanos, D. Llorens, A. Marzal, F. Prat, E. Vidal and J.M. Vilar: “Using Categories in the Eutrans System”. ACL-ELSNET Workshop on Spoken Language Translation, Madrid, Spain, pp. 44–52. (1997)Google Scholar
  2. 2.
    J.C. Amengual, E. Vidal. Two Different Approaches for Cost-efficient Viterbi Parsing with Error Correction. Proc. of the SSPR'96, IAPR International Workshop on Structural and Syntactical Pattern Recognition, August 20–23, 1996, Leipzig. To be published in the Proceedings.Google Scholar
  3. 3.
    J.C. Amengual, E. Vidal and J.M. Benedí. “Simplifying Language through Error-Correcting Decoding”. Proceedings of the ICSLP96 (IV International Conference on Spoken Language Processing). To be published. October, 1996.Google Scholar
  4. 4.
    J. Berstel. Transductions and Context-Free Languages. Teubner, Stuttgart. 1979.Google Scholar
  5. 5.
    P.F. Brown et al.. “A Statistical Approach to Machine Translation”. Computational Linguistics, Vol. 16, No.2, pp.79–85, 1990.Google Scholar
  6. 6.
    A. Castellanos, E. Vidal, J. Oncina. “Language Understanding and Subsequential Transducer Learning”. 1st International Colloquium on Grammatical Inference,Colchester, England. proc., pp. 11/1–11/10. April, 1993.Google Scholar
  7. 7.
    J.G. Bauer, H. Stahl, J. Mller: “A One-pass Search Algorithm for Understanding Natural Spoken Time Utterances by Stochastic Models”. Proc. of the EUROSPEECH'95, Madrid, Spain, vol.I, pp. 567–570. (1995)Google Scholar
  8. 8.
    F. Casacuberta: “Maximum Mutual Information and Conditional Maximum Likelihood Estimations on Stochastic Regular Syntax-Directed TranslationSchemes”, Lecture notes in Artificial Intelligence, vol.1147, pp. 282–291, Springer-Verlag. (1996)Google Scholar
  9. 9.
    F. Casacuberta: “Growth Transformations for Probabilistic Functions on Stochastic Grammars”, International Journal of Pattern Recognition and Artificial Intelligence, vol. 10, n. 3, pp. 183–201, Word Scientific Publishing Company. (1996)Google Scholar
  10. 10.
    C.T. Hemphill, J.J. Godfrey, G.R. Doddington. “The ATIS Spoken Language Systems, pilot Corpus”. Proc. of 3rd DARPA Workshop on Speech and Natural Language, pp. 102–108, Hidden Valley (PA), June 1990.Google Scholar
  11. 11.
    F. Jelinek: “Language Modeling for Speech Recognition”. In [13] (1996).Google Scholar
  12. 12.
    V. Jimenez, A. Castellanos, E. Vidal. “Some results with a trainable speech translation and understanding system”. In Proceedings of the ICASSP-95, Detroit, MI (USA), 1995Google Scholar
  13. 13.
    A. Kornai (ed.); Proceedings of the ECAI'96 Workshop: Extended Finite State Models of Language. Budapest, 1996.Google Scholar
  14. 14.
    A. Lavie, A. Waibel, L. Levin, M. Finke, D. Gates, M. Gavaldà, T. Zeppenfeld and P. Zhan: “JANUS-III: Speech-to-speech Translation in Multiple Languages”, Proc. of the ICASSP'97, Munich, Germany, vol.I, pp. 99–102. (1997)Google Scholar
  15. 15.
    E. Maier and S. McGlashan: “Semantic and Dialogue Processing in the VERBMOBIL Spoken Dialogue Translation System”, In Proceedings in Artificial Intelli-gence: CRIM/FORWISS Workshop on Progress and Prospects of Speech Research and Technology, H. Niemann, R. de Mori and G. Hanrieder (eds.), Infix, pp. 270–273. (1994)Google Scholar
  16. 16.
    J. Oncina. “Aprendizaje de Lenguages RegulÄres y Funciones Subsecuenciales”. Ph.D. diss., Universidad Politecnica de Valencia, 1991.Google Scholar
  17. 17.
    J. Oncina, P. Garcia, E. Vidal. “Learning Subsequential Transducers for Pattern Recognition Interpretation Tasks”. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.15, No.5, pp.448–458. May, 1993.CrossRefGoogle Scholar
  18. 18.
    J. Oncina, A. Castellanos, E. Vidal, V. Jimenez. “Corpus-Based Machine Translation through Subsequential Transducers”. Third Int. Conf. on the Cognitive Science of Natural Language Processing, proc., Dublin, 1994Google Scholar
  19. 19.
    J. Oncina, M.A. Var. “Using domain information during the learning of a subsequential transducer”. In Laurent Miclet and Colin de la Higuera, editors, Gram-matical Inference: Learning Syntax from Sentences, Lecture Notes in Computer Science, vol. 1147, pp. 301–312. Springer-Verlag. 1996Google Scholar
  20. 20.
    R. Pieraccini, E. Levin. “Stochastic Representation of Semantic Structure for Speech Understanding”. EUROSPEECH'91, Proc., Vol. 2, pp.383–386. Genoa Sept, 1991.Google Scholar
  21. 21.
    R. Pieraccini, E. Levin, E. Vidal. “Learning How To Understand Language”. EUROSPEECH'93, proc., Vol.2, pp. 1407–1412. Berlin, Sept, 1993.Google Scholar
  22. 22.
    N. Prieto, E. Vidal. “Learning Language Models through the ECGI method”. Speech Communication, No.11, pp.299–309. 1992.CrossRefGoogle Scholar
  23. 23.
    K. Seymore, R. Rosenfeld. “Scalable Backoff Language Models”. ICSLP-96, proc. pp.232–235. Philadelfia, 1996.Google Scholar
  24. 24.
    E. Vidal: “Language Learning, Understanding and Translation”, In Proc. in Art. Intell.: CRIM/FORWISS Workshop on Progress and Prospects of Speech Research and Technology, H. Niemann, R. de Mori and G. Hanrieder (eds.), pp. 131–140. Infix, (1994).Google Scholar
  25. 25.
    E. Vidal: “Finite-State Speech-to-speech Translation”, Proc. of the ICASSP'97,Munich, Germany, vol.1, pp. 111–122. (1997)Google Scholar
  26. 26.
    E. Vidal, F. Casacuberta, P. Garcia. “Grammatical Inference and Automatic Speech Recognition”. In Speech Recognition and Coding. New Advances and Trends, J.Rubio and J.M.Lopez, Eds. Springer Verlag, 1994.Google Scholar
  27. 27.
    E. Vidal, D. Llorens. “Using knowledge to improve N-Gram Language Modeling through the MGGI methodology”. In Grammatical Inference: Learning Syntax from Sentences, L.Miclet, C.De La Higuera, Eds. LNAI (1147), Springer-Verlag, 1996.Google Scholar
  28. 28.
    J.M. Vilar, A. Marzal, E. Vidal: “Learning Language Translation in Limited Domains using Finite-State Models: some Extensions and Improvements”. Proceedings of the EUROSPEECH-95, Madrid, Spain, pp. 1231–1234. (1995)Google Scholar
  29. 29.
    J.M. Vilar, E. Vidal and J.C. Amengual: “Learning Extended Finite State Models for Language Translation”. Proceedings of the ECAI96 (12th European Conference on Artificial Intelligence). August (1996).Google Scholar
  30. 30.
    Linux system documentacion, at directory “/usr/doc/at” (Debian distribution). Also, see “man at” on a Unix system.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1998

Authors and Affiliations

  • David Picó
    • 1
  • Enrique Vidal
    • 1
  1. 1.Institut TecnolÒgic d'InformàticaUniversitat Politècnica de ValènciaValènciaSpain

Personalised recommendations