Tokenization of Chinese input text into words is a necessary step to realize a Mandarin Chinese text-to-speech. Several word-segmentation algorithms were developed in which linguistic information are combined with statistical ones or with heuristic rules. In this paper we investigate in the advantages that can arise when semantic relation among sentences is taken into account during the word se...

We establish the following characteristics of the task of perspective classification: (a) using term frequencies in a document does not improve classification achieved with absence/presence features; (b) for datasets allowing the relevant comparisons, a small number of top features is found to be as effective as the full feature set and indispensable for the best achieved performance, testifyin...

In this study we investigate vowel reduction and the role of some lexical factors in the production of vowels extracted from a corpus of French conversations. Vowel durations and spectral quality are examined with respect to 1. their interaction in the corpus, 2. the position of vowels in words, and 3. word frequency and word category. The analyses are conducted on vowels produced by 16 speaker...

A frequently replicated finding is that higher frequency words tend to be shorter and contain more strongly reduced vowels. However, little is known about potential differences in the articulatory gestures for high vs. low frequency words. The present study made use of electromagnetic articulography to investigate the production of two German vowels, [i] and [a], embedded in high and low freque...

In spontaneous speech, speakers sometimes replace a word they have just produced or started producing by another word. The present study reports that in these replacement repairs, low-frequency replaced words are more likely to be interrupted prior to completion than high-frequency words, providing support to the hypothesis that the production of high-frequency words is more automatic than the ...