This study presents an N-gram adaptation technique when additional text data for the adaptation do not exist. We use a language modeling approach to the information retrieval (IR) technique to collect the appropriate adaptation corpus from baseline text data. We propose to use a dynamic interpolation coefficient to merge the N-gram, where the interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech. Experimental results show that the proposed adapted N-gram always has better performance than the background N-gram.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Joon-Ki CHOI, Yung-Hwan OH, "N-gram Adaptation with Dynamic Interpolation Coefficient Using Information Retrieval Technique" in IEICE TRANSACTIONS on Information,
vol. E89-D, no. 9, pp. 2579-2582, September 2006, doi: 10.1093/ietisy/e89-d.9.2579.
Abstract: This study presents an N-gram adaptation technique when additional text data for the adaptation do not exist. We use a language modeling approach to the information retrieval (IR) technique to collect the appropriate adaptation corpus from baseline text data. We propose to use a dynamic interpolation coefficient to merge the N-gram, where the interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech. Experimental results show that the proposed adapted N-gram always has better performance than the background N-gram.
URL: https://global.ieice.org/en_transactions/information/10.1093/ietisy/e89-d.9.2579/_p
Copy
@ARTICLE{e89-d_9_2579,
author={Joon-Ki CHOI, Yung-Hwan OH, },
journal={IEICE TRANSACTIONS on Information},
title={N-gram Adaptation with Dynamic Interpolation Coefficient Using Information Retrieval Technique},
year={2006},
volume={E89-D},
number={9},
pages={2579-2582},
abstract={This study presents an N-gram adaptation technique when additional text data for the adaptation do not exist. We use a language modeling approach to the information retrieval (IR) technique to collect the appropriate adaptation corpus from baseline text data. We propose to use a dynamic interpolation coefficient to merge the N-gram, where the interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech. Experimental results show that the proposed adapted N-gram always has better performance than the background N-gram.},
keywords={},
doi={10.1093/ietisy/e89-d.9.2579},
ISSN={1745-1361},
month={September},}
Copy
TY - JOUR
TI - N-gram Adaptation with Dynamic Interpolation Coefficient Using Information Retrieval Technique
T2 - IEICE TRANSACTIONS on Information
SP - 2579
EP - 2582
AU - Joon-Ki CHOI
AU - Yung-Hwan OH
PY - 2006
DO - 10.1093/ietisy/e89-d.9.2579
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E89-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2006
AB - This study presents an N-gram adaptation technique when additional text data for the adaptation do not exist. We use a language modeling approach to the information retrieval (IR) technique to collect the appropriate adaptation corpus from baseline text data. We propose to use a dynamic interpolation coefficient to merge the N-gram, where the interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech. Experimental results show that the proposed adapted N-gram always has better performance than the background N-gram.
ER -