Title: Robust Analysis of Spoken Input Combining Statistical and Knowledge-based Information Sources

Authors: Roldano Cattoni, Marcello Federico, Alon Lavie

Abstract: The work presented in this paper concerns the analysis of automatic transcription of spoken input into an interlingua formalism for a speech-to-speech machine translation system. This process is based on two sub-tasks, (1) the recognition of the speech act and possible concepts and (2) the extraction of possible feature-value information called arguments. Statistical models are used for the former, while a knowledge-based approach is employed for the latter. This paper proposes an algorithms that improves the analysis step in terms of robustness and performance: it integrates the scores of the statistical models with the extracted arguments, taking in account the a-priori constraints defined by the interlingua formalism.

a01rc059.ps a01rc059.pdf