List of Figures List of Tables 1 Analogies between Hmms and Stochastic Linear

In this paper we present a non-traditional approach to the problem of estimating the parameters of a stochastic linear system. The method is based on the ExpectationMaximization algorithm and can be considered as the continuous analog of the BaumWelch estimation algorithm for hidden Markov models. We use the algorithm for training the parameters of a dynamical system model that we propose for better representing the spectral dynamics of speech for recognition. We assume that the observed feature vectors of a phone segment are the output of a stochastic linear dynamical system and we show how the evolution of the dynamics as a function of the segment length can be modeled using alternative assumptions. We show on a phoneme classi cation task using the TIMIT database that our approach is the rst e ective use of an explicit model for statistical dependence between frames of speech.