The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS

Kong Aik Lee; Haizhou Li; Li Deng; Ville Hautamäki; Wei Rao; Xiong Xiao; Anthony Larcher; Hanwu Sun; Trung Hieu Nguyen; Guangsen Wang; Aleksandr Sizov; Jianshu Chen; Ivan Kukanov; Amir Hossein Poorjam; Trung Ngo Trong; Cheng Lin Xu; Hai Hua Xu; Bin Ma; Eng Siong Chng; Sylvain Meignier

doi:10.21437/Interspeech.2016-624

The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS

Kong Aik Lee, Haizhou Li, Li Deng, Ville Hautamäki, Wei Rao, Xiong Xiao, Anthony Larcher, Hanwu Sun, Trung Hieu Nguyen, Guangsen Wang, Aleksandr Sizov, Jianshu Chen, Ivan Kukanov, Amir Hossein Poorjam, Trung Ngo Trong, Cheng Lin Xu, Hai Hua Xu, Bin Ma, Eng Siong Chng, Sylvain Meignier

Research output: Journal article publication › Conference article › Academic research › peer-review

20 Citations (Scopus)

Abstract

The series of language recognition evaluations (LRE's) conducted by the National Institute of Standards and Technology (NIST) have been one of the driving forces in advancing spoken language recognition technology. This paper presents a shared view of five institutions resulting from our collaboration toward LRE 2015 submissions under the names of I2R, Fantastic4, and SingaMS. Among others, LRE'15 emphasizes on language detection in the context of closely related languages, which is different from previous LRE's. From the perspective of language recognition system design, we have witnessed a major paradigm shift in adopting deep neural network (DNN) for both feature extraction and classifier. In particular, deep bottleneck features (DBF) have a significant advantage in replacing the shifted-delta-cepstral (SDC) which has been the only option in the past. We foresee deep learning is going to serve as a major driving force in advancing spoken language recognition system in the coming years.

Original language	English
Pages (from-to)	3211-3215
Number of pages	5
Journal	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Volume	08-12-September-2016
DOIs	https://doi.org/10.21437/Interspeech.2016-624
Publication status	Published - Sept 2016
Externally published	Yes
Event	17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016 - San Francisco, United States Duration: 8 Sept 2016 → 16 Sept 2016

Keywords

Evaluation
Spoken language recognition

ASJC Scopus subject areas

Language and Linguistics
Human-Computer Interaction
Signal Processing
Software
Modelling and Simulation

More information

10.21437/Interspeech.2016-624

Cite this

Lee, K. A., Li, H., Deng, L., Hautamäki, V., Rao, W., Xiao, X., Larcher, A., Sun, H., Nguyen, T. H., Wang, G., Sizov, A., Chen, J., Kukanov, I., Poorjam, A. H., Trong, T. N., Xu, C. L., Xu, H. H., Ma, B., Chng, E. S., & Meignier, S. (2016). The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 08-12-September-2016, 3211-3215. https://doi.org/10.21437/Interspeech.2016-624

@article{620576fe762241b6b24c511d84947a8f,

title = "The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS",

abstract = "The series of language recognition evaluations (LRE's) conducted by the National Institute of Standards and Technology (NIST) have been one of the driving forces in advancing spoken language recognition technology. This paper presents a shared view of five institutions resulting from our collaboration toward LRE 2015 submissions under the names of I2R, Fantastic4, and SingaMS. Among others, LRE'15 emphasizes on language detection in the context of closely related languages, which is different from previous LRE's. From the perspective of language recognition system design, we have witnessed a major paradigm shift in adopting deep neural network (DNN) for both feature extraction and classifier. In particular, deep bottleneck features (DBF) have a significant advantage in replacing the shifted-delta-cepstral (SDC) which has been the only option in the past. We foresee deep learning is going to serve as a major driving force in advancing spoken language recognition system in the coming years.",

keywords = "Evaluation, Spoken language recognition",

author = "Lee, {Kong Aik} and Haizhou Li and Li Deng and Ville Hautam{\"a}ki and Wei Rao and Xiong Xiao and Anthony Larcher and Hanwu Sun and Nguyen, {Trung Hieu} and Guangsen Wang and Aleksandr Sizov and Jianshu Chen and Ivan Kukanov and Poorjam, {Amir Hossein} and Trong, {Trung Ngo} and Xu, {Cheng Lin} and Xu, {Hai Hua} and Bin Ma and Chng, {Eng Siong} and Sylvain Meignier",

note = "Publisher Copyright: Copyright {\textcopyright} 2016 ISCA.; 17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016 ; Conference date: 08-09-2016 Through 16-09-2016",

year = "2016",

month = sep,

doi = "10.21437/Interspeech.2016-624",

language = "English",

volume = "08-12-September-2016",

pages = "3211--3215",

journal = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

issn = "2308-457X",

}

Lee, KA, Li, H, Deng, L, Hautamäki, V, Rao, W, Xiao, X, Larcher, A, Sun, H, Nguyen, TH, Wang, G, Sizov, A, Chen, J, Kukanov, I, Poorjam, AH, Trong, TN, Xu, CL, Xu, HH, Ma, B, Chng, ES & Meignier, S 2016, 'The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS', Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, vol. 08-12-September-2016, pp. 3211-3215. https://doi.org/10.21437/Interspeech.2016-624

The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS. / Lee, Kong Aik; Li, Haizhou; Deng, Li et al.
In: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, Vol. 08-12-September-2016, 09.2016, p. 3211-3215.

Research output: Journal article publication › Conference article › Academic research › peer-review

TY - JOUR

T1 - The 2015 NIST language recognition evaluation

T2 - 17th Annual Conference of the International Speech Communication Association, INTERSPEECH 2016

AU - Lee, Kong Aik

AU - Li, Haizhou

AU - Deng, Li

AU - Hautamäki, Ville

AU - Rao, Wei

AU - Xiao, Xiong

AU - Larcher, Anthony

AU - Sun, Hanwu

AU - Nguyen, Trung Hieu

AU - Wang, Guangsen

AU - Sizov, Aleksandr

AU - Chen, Jianshu

AU - Kukanov, Ivan

AU - Poorjam, Amir Hossein

AU - Trong, Trung Ngo

AU - Xu, Cheng Lin

AU - Xu, Hai Hua

AU - Ma, Bin

AU - Chng, Eng Siong

AU - Meignier, Sylvain

PY - 2016/9

Y1 - 2016/9

N2 - The series of language recognition evaluations (LRE's) conducted by the National Institute of Standards and Technology (NIST) have been one of the driving forces in advancing spoken language recognition technology. This paper presents a shared view of five institutions resulting from our collaboration toward LRE 2015 submissions under the names of I2R, Fantastic4, and SingaMS. Among others, LRE'15 emphasizes on language detection in the context of closely related languages, which is different from previous LRE's. From the perspective of language recognition system design, we have witnessed a major paradigm shift in adopting deep neural network (DNN) for both feature extraction and classifier. In particular, deep bottleneck features (DBF) have a significant advantage in replacing the shifted-delta-cepstral (SDC) which has been the only option in the past. We foresee deep learning is going to serve as a major driving force in advancing spoken language recognition system in the coming years.

AB - The series of language recognition evaluations (LRE's) conducted by the National Institute of Standards and Technology (NIST) have been one of the driving forces in advancing spoken language recognition technology. This paper presents a shared view of five institutions resulting from our collaboration toward LRE 2015 submissions under the names of I2R, Fantastic4, and SingaMS. Among others, LRE'15 emphasizes on language detection in the context of closely related languages, which is different from previous LRE's. From the perspective of language recognition system design, we have witnessed a major paradigm shift in adopting deep neural network (DNN) for both feature extraction and classifier. In particular, deep bottleneck features (DBF) have a significant advantage in replacing the shifted-delta-cepstral (SDC) which has been the only option in the past. We foresee deep learning is going to serve as a major driving force in advancing spoken language recognition system in the coming years.

KW - Evaluation

KW - Spoken language recognition

UR - http://www.scopus.com/inward/record.url?scp=84994268371&partnerID=8YFLogxK

U2 - 10.21437/Interspeech.2016-624

DO - 10.21437/Interspeech.2016-624

M3 - Conference article

AN - SCOPUS:84994268371

SN - 2308-457X

VL - 08-12-September-2016

SP - 3211

EP - 3215

JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

Y2 - 8 September 2016 through 16 September 2016

ER -

The 2015 NIST language recognition evaluation: The shared view of I2R, fantastic4 and singaMS

Abstract

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this