 |
Conference papers
The 2015 NIST Language Recognition Evaluation: The Shared View of I2R, Fantastic4 and SingaMS
Kong Aik Lee
1
Haizhou Li
1
Li Deng
2
Ville Hautamäki
3
Wei Rao
4
Xiong Xiao
4
Anthony Larcher
5
Hanwu Sun
1
Trung Nguyen
1
Guangsen Wang
1
Aleksandr Sizov
1, 3
Jianshu Chen
2
Ivan Kukanov
3
Amir Poorjam
3
Trung Trong
3
Cheng-Lin Xu
4
Hai-Hua Xu
4
Bin Ma
1
Eng-Siong Chng
4
Sylvain Meignier
5
Wei Rao 4
Author
Xiong Xiao 4
Author
Anthony Larcher 5
Author IdHAL : anthony-larcher ORCID : https://orcid.org/0000-0003-4398-0224
Hanwu Sun 1
Author
Hai-Hua Xu 4
Author
Bin Ma 1
Author
Eng-Siong Chng 4
Author
Sylvain Meignier 5
Author IdHAL : sylvain-meignier arXiv : https://arxiv.org/a/sylvain.meignier ORCID : https://orcid.org/0000-0001-7687-073X
1
Networking Protocols Department (Agency for Science, Technology and Research (A*STAR), Institute for Infocomm Research (I2R)) (Singapore)
2
Microsoft Research [Redmond] (One Microsoft Way, Redmond, WA 98052, USA - United States)
- Microsoft Corporation [Redmond, Wash.] (United States)
3
University of Eastern Finland ( Educa, Tulliportinkatu 1, 80130 Joensuu, Finland - Finland)
4
Nanyang Technological University [Singapour] (50 Nanyang Ave, Singapour 639798 - Singapore)
5
LIUM - Laboratoire d'Informatique de l'Université du Mans (Avenue Laennec 72085 Le Mans cedex 9 - France)
- UM - Le Mans Université : EA4023 (Avenue Olivier Messiaen - 72085 Le Mans cedex 9 - France)
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, Fan-tastic4, 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.
Document type :
Conference papers
Complete list of metadata
https://hal.archives-ouvertes.fr/hal-01433164
Contributor : Sylvain Meignier <>
Submitted on : Friday, March 24, 2017 - 11:48:25 PM
Last modification on : Thursday, December 19, 2019 - 1:50:04 PM
Long-term archiving on: : Sunday, June 25, 2017 - 12:27:57 PM
Contributor : Sylvain Meignier <>
Submitted on : Friday, March 24, 2017 - 11:48:25 PM
Last modification on : Thursday, December 19, 2019 - 1:50:04 PM
Long-term archiving on: : Sunday, June 25, 2017 - 12:27:57 PM
File
0624_v2.pdf
Publisher files allowed on an open archive
Identifiers
- HAL Id : hal-01433164, version 1
- DOI : 10.21437/Interspeech.2016-624
Collections
UNIV-LEMANS | LIUM | LIUM-LST
Citation
Kong Aik Lee, Haizhou Li, Li Deng, Ville Hautamäki, Wei Rao, et al.. The 2015 NIST Language Recognition Evaluation: The Shared View of I2R, Fantastic4 and SingaMS. Interspeech 2016, 2016, San Fransisco, United States. pp.3211 - 3215, ⟨10.21437/Interspeech.2016-624⟩. ⟨hal-01433164⟩
Metrics
Record views
569
Files downloads