This study analyses how the reduction of the look-ahead length of a two pass phonetic decoder influences the alignment of the segment boundaries. It is shown how the optimization ...
Human listeners use lexical stress for word segmentation and disambiguation. We look into using lexical stress for speech recognition by examining a Dutch-language corpus. We propo...
Abstract. We are introducing several methods for combination of systems based on temporal trajectories feature-level combination. Experiments were done to evaluate combination meth...
In this paper we present an adaptive architecture for interaction and dialogue management in spoken dialogue applications. This architecture is targeted for applications that adapt...
This article considers the error of integrating multivariate Haar wavelet series by quasi-Monte Carlo rules using scrambled digital nets. Both the worst-case and random-case errors...
Stefan Heinrich, Fred J. Hickernell, Rong-Xian Yue