Some approaches to statistical and finite-state speech-to-speech translation F. Casacuberta a, * , H. Ney b , F.J. Och b , E. Vidal a , J.M. Vilar d , S. Barrachina c , I. Garc  ıa-Varea a , D. Llorens d , C. Mart  ınez a , S. Molau b , F. Nevado a , M. Pastor a , D. Pic o a , A. Sanchis a , C. Tillmann b a Departament de Sistemes Inform atics i Computaci o, Institut Tecnol ogic d’Inform atica, Universitat Polit ecnica de Val encia, Camino de Vera, s/n, Valencia 46071, Spain b Lehrstuhl f € ur Informatik VI, RWTH Aachen, University of Technology, Aachen D-52056, Germany c Departament d’Enginyeria i Ci encies dels Computadors, Universitat Jaume I, Castell o de la Plana 12071, Spain d Departament de Llenguatges i Sistemes Inform atics, Universitat Jaume I, Castell o de la Plana 12071, Spain Received 3 December 2001; received in revised form 14 April 2003; accepted 20 May 2003 Abstract Speech-input translation can be properly approached as a pattern recognition problem by means of statistical alignment models and stochastic finite-state transducers. Under this general framework, some specific models are presented. One of the features of such models is their capability of automatically learning from training examples. Moreover, the stochastic finite-state transducers permit an integrated architecture similar to one used in speech recognition. In this case, the acoustic models (hidden Markov models) are embedded into the finite-state transducers, and the translation of a source utterance is the result of a (Viterbi) search on the integrated network. These approaches have been followed in the framework of the European project EUTRANS RANS . Translation experiments have been performed from Spanish to English and from Italian to English in an application involving the interaction of a customer with a receptionist at the frontdesk of a hotel. Ó 2003 Elsevier Ltd. All rights reserved. Keywords: Speech-input translation; Statistical alignment models; Stochastic finite-state transducers * Corresponding author. Tel.: +34-96-387-7241; fax: +34-96-387-7239. E-mail address: fcn@iti.upv.es (F. Casacuberta). 0885-2308/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0885-2308(03)00028-7 Computer Speech and Language 18 (2004) 25–47 COMPUTER SPEECH AND LANGUAGE www.elsevier.com/locate/csl