David Yarowsky. Homograph disambiguation in text-to-speech synthesis. In Jan van Santen, Richard Sproat, Joseph Olive, and Julia Hirschberg, editors, Progress in Speech Synthesis, pages 157-172. Springer, New York,  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... cases include dates in month day or month year format (e.g. 1 2, for January the second) which are systematically ambiguous with fractions (one half) and three or four digit numbers which are system atically ambiguous between dates and ordinary number names (in 1901, 1901 tons) Yarowsky [30] demonstrated good performance on disambiguating such cas es using decision list based techniques, which had previously been developed for more general sense disambiguation problems. Once again though, such tech niques do presume that you know beforehand the individual cases that must be ....

....cases (e.g. St for Saint versus Street) that are better cast in terms of features of the context. Class based language models and feature based back off mechanisms [3] offer possible solutions for the approach based on a source channel model. Alternatively, previous work such as that of Yarowsky [30] applying decision lists to particular instances of these problems suggests that direct models might be an effective solution, which would likely require maximum entropy techniques. For the unsupervised training paradigm, it may be that cache based language models are useful. As we have hinted at ....

David Yarowsky. Homograph disambiguation in text-to-speech synthesis. In Jan van Santen, Richard Sproat, Joseph Olive, and Julia Hirschberg, editors, Progress in Speech Synthesis, pages 157-172. Springer, New York,

....segmentation or numeral expansion in a preprocessing phase that is logically prior to the linguistic analysis phase; we have argued here against this view. Areas of future work include incorporating decision tree based models of phrasing [13] and decision list based sense disambiguation methods [14] into the text analysis model using the tree compiler described in [10] and similar tools. 6. ....

David Yarowsky. Homograph disambiguation in text-to-speech synthesis. In Jan van Santen, Richard Sproat, Joseph Olive, and Julia Hirschberg, editors, Progress in Speech Synthesis. Springer, New York, 1996.

....methods previously described will return a set of pronunciations though often a set with a single member. In cases where there is more than one viable candidate, one must have some method for choosing between them, a problem referred to as homograph disambiguation. One approach is described in [Yarowsky, 1997]. One starts with a training corpus containing tagged 19 examples in context of each pronunciation of a homograph. Significant evidence from both narrow and wide context (e.g. n gram word patterns containing the homograph in question or words that occur anywhere in the same sentence, which are ....

Yarowsky, D. (1997). Homograph disambiguation in text-to-speech synthesis. In SSOH, editor, Progress in Speech Synthesis, pages 157--172. Springer, New York, NY.

....Japanese, one would consider the means by which abbreviations differ from their intended word and identify features to capture these differences. Related Research The majority of related research has occurred in the field of text to speech synthesis as a part of the process of text normalization (Yarowsky, 1996 ; Allen et al., 1987 ; Sproat, 1998) The goal of text normalization is to convert written text into a format that is suitable for speech generation. Hence, abbreviations must be expanded, digits must be identified as to whether they are phone numbers, post codes, money amounts, and so on. ....

....that the problem is one of disambiguation rather than expansion. That is, it is assumed that one knows that a given word is an abbreviation and that one also knows what the possible candidate words are. The problem is then to determine which of the candidate words is intended in a particular case (Yarowsky, 1996 ; Sproat, 1998) However, for many applications it is impossible to develop a priori a list of abbreviations and their possible expansions. In the case of the ASRS database, each report is authored by a different individual, there is no predicting which words will be abbreviated or how they will ....

Yarowsky, D. (1996). Homograph disambiguation in text-to-speech synthesis. In J. van Santen, R. Sproat, J. Olive & J. Hirschberg (Eds). Progress in Speech Synthesis. New York, Springer: 157-172. .

.... Word Word to right PoS Function Word Two words to left PoS Two words to right PoS Word to right and left PoS Both Function Words Global content word collocations Word in Window of 4 Word in sentence Content Word Table 3: Kinds of collocations considered and homograph disambiguation (Yarowsky, 1994; 1995; 1996). In order to build decision lists the training examples are processed to extract the features (each feature corresponds to a kind of collocation) which are weighted with a log likelihood measure. The list of all features ordered by log likelihood values constitutes the decision list. We adapted ....

Yarowsky, D. Homograph Disambiguation in Textto -speech Synthesis. J Hirschburg, R. Sproat and J. Van Santen (eds.) Progress in Speech Synthesis, Springer-Vorlag, pp. 159-175. 1996.

....training. Finally some conclusions are drawn. 1 Decision lists and the features used Decision lists (DL) as defined in (Yarowsky, 1994) are simple means to solve ambiguity problems. They have been successfully applied to accent restoration, word sense disambiguation and homograph disambiguation (Yarowsky, 1994; 1995; 1996). It was one of the most successful systems on the Senseval word sense disambiguation competition (Kilgarriff and Palmer, 2000) The training data is processed to extract the features, which are weighted with a log likelihood measure. The list of all features ordered by the log likelihood values ....

Yarowsky, D. Homograph Disambiguation in Text-tospeech Synthesis. J Hirschburg, R. Sproat and J. Van Santen (eds.) Progress in Speech Synthesis, Springer-Vorlag, pp. 159-175. 1996.

....codes and the ignoring of graphical objects. Preprocessed text is then tokenized into word sized units that are looked up in the lexical database. A disambiguation module (Karaali et al. forthcoming) assigns words parts of speech (e.g. Church 1988) and in some cases, semantic information (e.g. Yarowsky 1997), to help in selecting the proper pronunciation for homographs. If a given word s pronunciation is not present in the dictionary at all, a pronunciation is generated for that word by the letter to sound conversion module. Once lexical pronunciations have been determined for all the words of a ....

....lexical pronunciation database has almost 200,000 pronunciations, including over 1000 of which require disambiguation by part of speech, and over 200 of which require disambiguation by sense. Synthesizers have modules which can tag incoming words for part of speech (Church 1989) or 76 semantics (Yarowsky 1997) in order to help select the right pronunciations. Of course, this requires that lexica tag non homophonous homographs for part of speech or semantics, if required. One of the purposes of combining the three source lexica was to identify and tag for part of speech or semantics as many ....

Yarowsky, David. 1997. Homograph disambiguation in text-to-speech synthesis. In Progress in speech synthesis, ed. Jan P. H. van Santen, Richard W. Sproat, Joseph P. Olive, and Julia Hirschberg, 157-172.

....on the many features of language, text and speech. The collection and ordering of these predicates attempts to capture the many ways that a target may prime for a stimulus, and the many feature combinatorics that contribute to retrieval. It is somewhat the inverse of Yarowsky et al. s [SHY92, Yar96, Yar94] decision list algorithm for identifying homographs. In their work, the test that yields the highest score determines the pronunciation of the word or phrase. It is a best match approach. For Loq comparisons, the particular features that match are not as important as whether the cue and ....

David Yarowsky. Homograph Disambiguation in Text-to-Speech Synthesis. In Jan P.H. Van Santen, Richard W. Sproat, Joseph P. Olive, and Julia Hirschberg, editors, Progress in Speech Synthesis, chapter 12, pages 157--172. Springer, 1996.

No context found.

David Yarowsky. Homograph Disambiguation in Text-to-Speech Synthesis. In Progress in Speech Synthesis (Jan van Santen, Richard Sproat, Joseph Olive, and Julia Hirschberg, editors), pp. 159--174, New York: Springer, 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC