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

facebooktwittermailshare

Exploiting noisy web data by OOV ranking for low-resource keyword search

Abstract: 

Spoken keyword search in low-resource condition suffers from out-of-vocabulary (OOV) problem and insufficient text data for language model (LM) training. Web-crawled text data is used to expand vocabulary and to augment language model. However, the mismatching between web text and the target speech data brings difficulties to effective utilization. New words from web data need an evaluation to exclude noisy words or introduce proper probabilities. In this paper, several criteria to rank new words from web data are investigated and are used as features
for logistic regression. In the IV keyword case, top N words are selected to expand the vocabulary. In the OOV keyword case, all words are used for expansion but unigram probabilities are re-assigned by Zipf’s law. On Swahili keyword search, after further text filtering and LM interpolation, this strategy is observed to outperform a strong and commonly used baseline method for data selection.

up
0 users have voted:

Paper Details

Authors:
Ji Wu
Submitted On:
15 October 2016 - 7:55am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Zhipeng Chen
Paper Code:
P3-11
Document Year:
2016
Cite

Document Files

ISCSLP2016_Poster_Exploiting.pdf

(416)

Subscribe

[1] Ji Wu, "Exploiting noisy web data by OOV ranking for low-resource keyword search", IEEE SigPort, 2016. [Online]. Available: http://sigport.org/1232. Accessed: Feb. 27, 2020.
@article{1232-16,
url = {http://sigport.org/1232},
author = {Ji Wu },
publisher = {IEEE SigPort},
title = {Exploiting noisy web data by OOV ranking for low-resource keyword search},
year = {2016} }
TY - EJOUR
T1 - Exploiting noisy web data by OOV ranking for low-resource keyword search
AU - Ji Wu
PY - 2016
PB - IEEE SigPort
UR - http://sigport.org/1232
ER -
Ji Wu. (2016). Exploiting noisy web data by OOV ranking for low-resource keyword search. IEEE SigPort. http://sigport.org/1232
Ji Wu, 2016. Exploiting noisy web data by OOV ranking for low-resource keyword search. Available at: http://sigport.org/1232.
Ji Wu. (2016). "Exploiting noisy web data by OOV ranking for low-resource keyword search." Web.
1. Ji Wu. Exploiting noisy web data by OOV ranking for low-resource keyword search [Internet]. IEEE SigPort; 2016. Available from : http://sigport.org/1232