Several characteristics of the Thai writing system make Thai grapheme-to-phoneme (G2P) conversion very challenging. In this paper, we propose an Example-Based Grapheme-toPhoneme conversion approach. It generates the pronunciation of a word by selecting, modifying and combining pronunciations from syllables from training corpus. The best system achieves 80.99% word accuracy and 94.19% phone accu...

We investigated the applicability of probabilistic context-free grammars to syllabi cation and grapheme-to-phoneme conversion. The results show that the standard probability model of context-free grammars performs very well in predicting syllable boundaries. However, our results indicate that the standard probability model does not solve grapheme-to-phoneme conversion su ciently although, we va...

We address a key problem in grapheme-tophoneme conversion: the ambiguity in mapping grapheme units to phonemes. Rather than using single letters and phonemes as units, we propose learning chunks, or subwords, to reduce ambiguity. This can be interpreted as learning a lexicon of subwords that has minimum description length. We implement an algorithm to build such a lexicon, as well as a simple d...

An Italian patient with a pure dysgraphia who incorrectly spelled words and nonwords is described. The spelling errors made by the patient were not affected by lexical factors (e.g., frequency, form class) and were qualitatively the same for words and nonwords. The pattern of writing performance is discussed in relation to current models of writing and, specifically, in relation to the role of ...

In synesthesia, certain stimuli ("inducers") may give rise to perceptual experience in additional modalities not normally associated with them ("concurrent"). For example, color-grapheme synesthetes automatically perceive achromatic numbers as colored (e.g., 7 is turquoise). Although synesthetes know when a given color matches the one evoked by a certain number, colors do not automatically give...

Synesthesia provides an elegant model to investigate neural mechanisms underlying individual differences in subjective experience in humans. In grapheme-color synesthesia, written letters induce color sensations, accompanied by activation of color area V4. Competing hypotheses suggest that enhanced V4 activity during synesthesia is either induced by direct bottom-up cross-activation from graphe...

Grapheme-color synaesthesia is a neurological phenomenon in which particular graphemes, such as the numeral 9, automatically induce the simultaneous perception of a particular color, such as the color red. To test whether the concurrent color sensations in grapheme-color synaesthesia are treated as meaningful stimuli, we recorded event-related brain potentials as 8 synaesthetes and 8 matched co...

Mirandese is the second official language in Portugal. For ages it was only preserved as an oral language and it was the object of a recent orthographic convention. This paper describes our efforts in porting our grapheme-to-phone module from European Portuguese to Mirandese. We describe the main differences between the two languages that can affect this module and the set of new SAMPA symbols ...