Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Phoneme Level Language Models for Sequence Based Low Resource ASR

Abstract: 

Building multilingual and crosslingual models help bring different languages together in a language universal space. It allows models to share parameters and transfer knowledge across languages, enabling faster and better adaptation to a new language. These approaches are particularly useful for low resource languages. In this paper, we propose a phoneme-level language model that can be used multilingually and for crosslingual adaptation to a target language. We show that our model performs almost as well as the monolingual models by using six times fewer parameters, and is capable of better adaptation to languages not seen during training in a low resource scenario. We show that these phoneme-level language models can be used to decode sequence based Connectionist Temporal Classification (CTC) acoustic model outputs to obtain comparable word error rates with Weighted Finite State Transducer (WFST) based decoding in Babel languages. We also show that these phoneme-level language models outperform WFST decoding in various low-resource conditions like adapting to a new language and domain mismatch between training and testing data.

up
0 users have voted:

Paper Details

Authors:
Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze
Submitted On:
14 May 2019 - 10:39am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Siddharth Dalmia
Paper Code:
4978
Document Year:
2019
Cite

Document Files

PLMs_ICASSP_Poster (1).pdf

(6)

Subscribe

[1] Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze, "Phoneme Level Language Models for Sequence Based Low Resource ASR", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4511. Accessed: May. 23, 2019.
@article{4511-19,
url = {http://sigport.org/4511},
author = {Siddharth Dalmia; Xinjian Li; Alan W Black; Florian Metze },
publisher = {IEEE SigPort},
title = {Phoneme Level Language Models for Sequence Based Low Resource ASR},
year = {2019} }
TY - EJOUR
T1 - Phoneme Level Language Models for Sequence Based Low Resource ASR
AU - Siddharth Dalmia; Xinjian Li; Alan W Black; Florian Metze
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4511
ER -
Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze. (2019). Phoneme Level Language Models for Sequence Based Low Resource ASR. IEEE SigPort. http://sigport.org/4511
Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze, 2019. Phoneme Level Language Models for Sequence Based Low Resource ASR. Available at: http://sigport.org/4511.
Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze. (2019). "Phoneme Level Language Models for Sequence Based Low Resource ASR." Web.
1. Siddharth Dalmia, Xinjian Li, Alan W Black, Florian Metze. Phoneme Level Language Models for Sequence Based Low Resource ASR [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4511