Ekonomiaren Garapen eta Lehiakortasun Saila

Morphological information is traditionally used to develop high quality text to speech (TTS) and automatic speech recognition (ASR) systems. The use of this information improves the naturalness and intelligibility of the TTS synthesis and provides an appropriated way to select lexical units (LU) for ASR. Basque is an agglutinative language with a complex structure inside the words and the morphological information is essential both in TTS and ASR. In this work an automatic morphological segmentation tool oriented to TTS and ASR tasks is presented.

An alternative to periodic boundary conditions is developed and tested in Monte Carlo simulations of the two- and three-dimensional Ising models. The boundary conditions are based on a mean-field approach that incorporates consistency constraints for the magnetization and correlations between nearest neighbors by means of an effective field and an extra coupling between nearest neighbors at the boundary of the simulation box. During the simulation the self-consistent equations are solved, and statistics are accumulated to obtain thermodynamic averages. In comparison with the standard periodic boundary conditions the method gives a more accurate estimation of nonuniversal magnitudes, such as the transition temperature and the behavior of the magnetization, but it cannot compete with the accuracy of other strategies such as finite-size scaling theory or Monte Carlo renormalization group to obtain critical exponents.