2012
DOI: 10.1109/tasl.2011.2167610
|View full text |Cite
|
Sign up to set email alerts
|

Experiments on Cross-Language Attribute Detection and Phone Recognition With Minimal Target-Specific Training Data

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
47
0

Year Published

2013
2013
2023
2023

Publication Types

Select...
5
2
2

Relationship

1
8

Authors

Journals

citations
Cited by 71 publications
(48 citation statements)
references
References 33 publications
1
47
0
Order By: Relevance
“…Those attributes could be identified from a particular language and shared across many different languages, so they could also be used to derive a universal set of speech units. Furthermore, data-sharing across languages at the acoustic phonetic attribute level is naturally facilitated by using these attributes, so more reliable language-independent acoustic parameter estimation can be anticipated [21]. In [16], it was also shown that these attributes can be used to compactly characterize any spoken language along the same lines as in the automatic speech attribute transcription (ASAT) paradigm for automatic speech recognition (ASR) [20].…”
Section: Choice and Extraction Of Attribute Featuresmentioning
confidence: 99%
See 1 more Smart Citation
“…Those attributes could be identified from a particular language and shared across many different languages, so they could also be used to derive a universal set of speech units. Furthermore, data-sharing across languages at the acoustic phonetic attribute level is naturally facilitated by using these attributes, so more reliable language-independent acoustic parameter estimation can be anticipated [21]. In [16], it was also shown that these attributes can be used to compactly characterize any spoken language along the same lines as in the automatic speech attribute transcription (ASAT) paradigm for automatic speech recognition (ASR) [20].…”
Section: Choice and Extraction Of Attribute Featuresmentioning
confidence: 99%
“…To avoid over-fitting, the reduction in classification error on the development set was adopted as the stopping criterion. The attribute detectors employed in this work were actually just those used in [21].…”
Section: Attribute Detector Designmentioning
confidence: 99%
“…The attribute questions performed a similar role to the articulatory features in e.g. [30] but without the explicit attribute detectors of these systems. The same phone symbol could correspond to different realisations of the associated phone across languages, especially when co-articulations are taken into account.…”
Section: Vowelmentioning
confidence: 93%
“…the Government Phonology (GP) [15,16], 2. the Sound Pattern of English (SPE) [17], and 3. the extended SPE system (eSPE) [18,19], Implementation of PhonVoc depends on the SSP vocoder and the Kaldi speech recognition toolkit [20] that we used for DNN training of both analysis and synthesis modules.…”
Section: Methodsmentioning
confidence: 99%