| Burget, L., Matějka, P., Černocký, J.: Discriminative Training Techniques for Acoustic Language Identification, In: Proceedings of ICASSP 2006, Toulouse, FR, 2006, p. 209-212 | | Publication language: | english |
|---|
| Original title: | Discriminative Training Techniques for Acoustic Language Identification |
|---|
| Title (cs): | Diskriminativní trénování pro automatickou identifikaci jazyka |
|---|
| Pages: | 209-212 |
|---|
| Proceedings: | Proceedings of ICASSP 2006 |
|---|
| Conference: | International Conference on Acoustics, Speech, and Signal Processing |
|---|
| Place: | Toulouse, FR |
|---|
| Year: | 2006 |
|---|
| URL: | http://www.fit.vutbr.cz/research/groups/speech/publi/2006/burget_mmi_lid_icassp2006.pdf [PDF] |
|---|
| Keywords |
|---|
language identification, language recognition, acoustic modeling, disriminative training, maximum mutual information
|
| Annotation |
|---|
This paper presents comparison of Maximum Likelihood (ML)
and discriminative Maximum Mutual Information (MMI) training
for acoustic modeling in language identification (LID).
Clear advantage of MMI over ML training is shown. The final error rate
compares favorably to other results published on NIST 2003 data. |
| Abstract |
|---|
This paper presents comparison of Maximum Likelihood (ML)
and discriminative Maximum Mutual Information (MMI) training
for acoustic modeling in language identification (LID). Both
approaches are compared on state-of-the-art shifted delta-cepstra
features, the results are reported on data from NIST 2003 evaluations.
Clear advantage of MMI over ML training is shown. Further
improvements of acoustic LID are discussed: Heteroscedastic Linear
Discriminant Analysis (HLDA) for feature de-correlation and
dimensionality reduction and Ergodic Hidden Markov models (EHMM) for
better modeling of dynamics in the acoustic space. The final error rate
compares favorably to other results published on NIST 2003 data.
|
| BibTeX: |
|---|
@INPROCEEDINGS{
author = {Lukáš Burget and Pavel Matějka and Jan Černocký},
title = {Discriminative Training Techniques for Acoustic Language
Identification},
pages = {209--212},
booktitle = {Proceedings of ICASSP 2006},
year = {2006},
location = {Toulouse, FR},
language = {english},
url = {http://www.fit.vutbr.cz/research/view_pub.php?id=8132}
} |
|