## Efficient Multilingual Phoneme-to-Grapheme Conversion Based on HMM (1996)

### Cached

### Download Links

- [acl.ldc.upenn.edu]
- [ucrel.lancs.ac.uk]
- [newdesign.aclweb.org]
- [wing.comp.nus.edu.sg]
- [aclweb.org]
- [www.aclweb.org]
- [aclweb.org]
- [www.aclweb.org]
- DBLP

### Other Repositories/Bibliography

Venue: | Computational Linguistics |

Citations: | 7 - 1 self |

### BibTeX

@ARTICLE{Rentzepopoulos96efficientmultilingual,

author = {Panagiotis A. Rentzepopoulos and George K. Kokkinakis},

title = {Efficient Multilingual Phoneme-to-Grapheme Conversion Based on HMM},

journal = {Computational Linguistics},

year = {1996},

volume = {22},

pages = {22--3}

}

### Years of Citing Articles

### OpenURL

### Abstract

Grapheme-to-phoneme conversion (GTPC) has been achieved in most European languagesby dictionary look-up or using rules. The application of these methods, however, in the reverse pro-cess, (i.e., in phoneme-to-grapheme conversion [PTGC]) creates serious problems, especially in inflectionally rich languages. In this paper the PTGC problem is approached from a completely different point of view. Instead of rules or a dictionary, the statistics of language connecting pro-nunciation to spelling are exploited. The novelty lies in modeling the natural language intraword features using the theory of hidden Markov models (HMM) and performing the conversion using the Viterbi algorithm. The PTGC system has been established and tested on various multilingual corpora. Initially, the first-order HMM and the common Viterbi algorithm were used to obtain a single transcription for each word. Afterwards, the second-order HMM and the N-best algorithm adapted to PTGC were implemented to provide one or more transcriptions for each word input (homophones). This system gave an average score of more than 99 % correctly transcribed words (overall success in the first four candidates)for most of the seven languages it was tested on (Dutch, English, French, German, Greek, Italian, and Spanish). The system can be adapted to almost any language with little effort and can be implemented in hardware to serve in real-time speech recognition systems. 1.

### Citations

4275 | A tutorial on hidden Markov models and selected applications in speech recognition
- Rabiner
- 1989
(Show Context)
Citation Context ...2.1 The First Order Hidden Markov Model An HMM can model any real-world process that changes states in time, provided that the state changes are more or less time independent (Hannaford and Lee 1990; =-=Rabiner 1989-=-). An HMM is used to describe statistical phenomena that can be considered sequences of hidden (i.e., not directly observable) states that produce observable symbols (Lee 1989). These phenomena are ca... |

747 | The Viterbi algorithm
- Forney
- 1973
(Show Context)
Citation Context ...ce Q(t) (graphemes) that maximizes the probability P(O I Q, ,~). A formal technique for finding the single best state sequence is based on dynamic programming and is the well-known Viterbi algorithm (=-=Forney 1973-=-; Viterbi 1967). In a word-level implementation, the algorithm must find the hidden-state sequence (i.e., word in its orthographic form) with the best score, given the model & and the observation sequ... |

664 | Slava M. Estimation of Probabilities from Sparse Data for the Language Model Component of a Speech Recognizer
- Katz
- 1987
(Show Context)
Citation Context ...On the other hand, if the training data are insufficient (something that would result in a very sparse transition matrix) then a smoothing technique should be used for the estimation function n' (x) (=-=Katz 1987-=-; Ney and Essen 1991). 2.2 Pilot System To implement the above algorithm in PTGC, some decisions had to be made about the states, observation symbols, and transition probabilities. These decisions are... |

63 | A comparison of several approximate algorithms for nding multiple (N-best) sentence hypotheses - Schwartz, Austin - 1991 |

13 |
Multi-Dimensional Hidden Markov Model of Telemanipulation Tasks With Varying Outcomes
- Hannaford, Lee
- 1990
(Show Context)
Citation Context ...real-time applications. 2.1 The First Order Hidden Markov Model An HMM can model any real-world process that changes states in time, provided that the state changes are more or less time independent (=-=Hannaford and Lee 1990-=-; Rabiner 1989). An HMM is used to describe statistical phenomena that can be considered sequences of hidden (i.e., not directly observable) states that produce observable symbols (Lee 1989). These ph... |

10 | Speaker Independent Phonetic Transcription of Fluent Speech for Large Vocabulary Speech Recognition
- Levinson, Liberman, et al.
- 1989
(Show Context)
Citation Context ...compare the phonemic strings to a (usually applicationspecific) dictionary containing both the phonemic and the graphemic form of every word the system can handle (Laface, Micca, and Pieraccini 1987; =-=Levinson et al. 1989-=-, etc.). Considering the effort and cost required to create such a dictionary, this is a serious limitation, especially for inflectionally rich languages such as Greek and German. Another very importa... |

8 | Some improvements in speech recognition algorithms based on hmm - Kriouile, Mari, et al. - 1990 |

8 |
A bi-directional model of English pronunciation
- Parfitt, Sharman
- 1991
(Show Context)
Citation Context ...ds such as "whatsoever", "therefore", etc.); this leads to many errors for an algorithm like the one proposed here, which has no information about the origin and etymology of each word. Similar work (=-=Parfitt and Sharman 1991-=-) shows the same problems in a slightly different context. Of course, more training of the model would improve performance. With French, there is a special problem, which does not occur with other lan... |

6 |
Hidden Markov Models: Past, Present, and Future
- Lee
- 1989
(Show Context)
Citation Context ...naford and Lee 1990; Rabiner 1989). An HMM is used to describe statistical phenomena that can be considered sequences of hidden (i.e., not directly observable) states that produce observable symbols (=-=Lee 1989-=-). These phenomena are called hidden Markov processes. A hidden Markov process is described by a model ,~ that consists of three matrices A, B, and ~-. Matrix A contains the transition probabilities o... |

6 |
Phoneme to Grapheme Conversion Using HMM
- Rentzepopoulos, Kokkinakis
- 1991
(Show Context)
Citation Context ... results of its evaluation, detailed in Section 3, were promising. For Greek, this system gave an average score of 78% correctly transcribed words, while at the phoneme level the score reached 95.5% (=-=Rentzepopoulos and Kokkinakis 1991-=-). Similar rates were achieved in four other languages (English, French, German, and Italian) (Rentzepopoulos and Kokkinakis 1992). The model implemented as above showed some disadvantages: • It did n... |

5 | Extended viterbi algorithm for second order hidden markov process - He - 1988 |

2 | Experimental results on a large lexicon access task - Laface, Pieraceini - 1987 |

2 |
Modern Greek Grammar and Comparative Analysis (in Greek
- Petrounias
- 1984
(Show Context)
Citation Context ..., and every phoneme usually has more than one possible spellings regardless of its neighboring phonemes. As an example, the phoneme/i/can be transcribed as z, 7/, v, ¢z, and oz in almost any context (=-=Petrounias 1984-=-; Setatos 1974). Other problems arise from the consonants, which can be either single or double without any change in the pronunciation. Finally, the model gave extremely good results with Italian and... |

1 | Linguistic analysis of the European languages. Technical Annex. European Commission Framework Programme "European Strategic Programme for Research in Information Technology - Project - 1987 |

1 |
European Commission Framework Programme "Linguistic Research and Engineering
- Ney, Essen
- 1991
(Show Context)
Citation Context ...r hand, if the training data are insufficient (something that would result in a very sparse transition matrix) then a smoothing technique should be used for the estimation function n' (x) (Katz 1987; =-=Ney and Essen 1991-=-). 2.2 Pilot System To implement the above algorithm in PTGC, some decisions had to be made about the states, observation symbols, and transition probabilities. These decisions are listed below. a. b.... |

1 |
Phoneme to Grapheme Conversion Using Rules (in Greek). Electrical Engineering Diploma thesis
- Rentzepopoulos
- 1988
(Show Context)
Citation Context ...raphically correct words; "#i),c~" 'speak!' and "#~)~c~" 'apples.' Previous work has shown that an average of 30 graphernic candidates is produced by this transcription for every input phonemic word (=-=Rentzepopoulos 1988-=-). To overcome the disadvantages of the above mentioned methods, a novel statistical approach to the problem of PTGC, which is based on hidden Markov models (HMM), has been investigated and is present... |

1 |
Multilingual phoneme to grapheme conversion system based on HMM
- Rentzepopoulos, Kokkinakis
- 1992
(Show Context)
Citation Context ...transcribed words, while at the phoneme level the score reached 95.5% (Rentzepopoulos and Kokkinakis 1991). Similar rates were achieved in four other languages (English, French, German, and Italian) (=-=Rentzepopoulos and Kokkinakis 1992-=-). The model implemented as above showed some disadvantages: • It did not have enough detail. • It could not produce more than one solution (homophones). Therefore, a higher-order HMM and a multiple-o... |

1 | Computational Linguistics Volume 22, Number 3 The N-best algorithm: An efficient and exact procedure for finding the N most likely sentence hypotheses - Schwartz, Chow - 1990 |

1 |
Phonology of the Modern Greek Koini (in Greek). Edited by Papazisis (editor
- Setatos
- 1974
(Show Context)
Citation Context ...me usually has more than one possible spellings regardless of its neighboring phonemes. As an example, the phoneme/i/can be transcribed as z, 7/, v, ¢z, and oz in almost any context (Petrounias 1984; =-=Setatos 1974-=-). Other problems arise from the consonants, which can be either single or double without any change in the pronunciation. Finally, the model gave extremely good results with Italian and Spanish, reac... |