Paper Info
Title
Models of tone for tonal and non-tonal languages.
Abstract
Conventional wisdom in automatic speech recognition asserts that pitch information is not helpful in building speech recognizers for non-tonal languages and contributes only modestly to performance in speech recognizers for tonal languages. To maintain consistency between different systems, pitch is therefore often ignored, trading the slight performance benefits for greater system uniformity/simplicity. In this paper, we report results that challenge this conventional approach. We present new models of tone that deliver consistent performance improvements for tonal languages (Cantonese, Vietnamese) and even modest improvements for non-tonal languages. Using neural networks for feature integration and fusion, these models achieve significant gains throughout, and provide us with system uniformity and standardization across all languages, tonal and non-tonal.
Year
DOI
Venue
2013
10.1109/ASRU.2013.6707740
ASRU
Keywords
Field
DocType
natural language processing,neural nets,speech recognition,automatic speech recognition,feature fusion,feature integration,neural network,nontonal language,pitch information,speech recognizer,tonal language,Acoustic Modeling,Automatic Speech Recognition,Neural Networks,Tonal Features,Tone Modeling
Speech corpus,Speech processing,Voice activity detection,Computer science,Speech recognition,Conventional wisdom,Artificial intelligence,Natural language processing,Artificial neural network,Standardization,Speech technology,Acoustic model
Conference
Citations 
PageRank 
References 
16
0.84
11
Authors
7
Name
Order
Citations
PageRank
Florian Metze11069106.49
Zaid A. W. Sheikh2203.62
Alex Waibel363431980.68
Jonas Gehring41487.87
Kevin Kilgour55411.00
Quoc Bao Nguyen6304.67
Van Huy Nguyen7181.92