Performance Comparison between Human Engineered Machine Learned Letter-to-Sound Rules for English: A Machine Learning Success Story (1993)
| Citations: | 1 - 0 self |
BibTeX
@MISC{Bakiri93performancecomparison,
author = {Ghulum Bakiri and Thomas G. Dietterich},
title = {Performance Comparison between Human Engineered Machine Learned Letter-to-Sound Rules for English: A Machine Learning Success Story},
year = {1993}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
The task of mapping spelled English words into strings of phonemes and stresses ("reading aloud") has many practical applications. Several commercial systems perform this task by applying a knowledge base of expert-supplied letter-to-sound rules. This paper presents a set of machine learning methods for automatically constructing letter-to-sound rules by analyzing a dictionary of words and their pronunciations. Our results demonstrate that these methods, taken together, provide a substantial performance improvement over the best commercial system---DECtalk from Digital Equipment Corporation----that utilizes hand-crafted letter-to-sound rules. The results are significant since these machine learning techniques are general and can be applied to many other tasks including the task of discovering pronunciation rules in other languages, there-by eliminating the need for hand-crafting a rule-base for each language. 1 Overview The automatic conversion of English text to synthetic speech is ...
