Publication Details

Progress in the BBN Keyword Search System for the DARPA RATS Program

NG Tim, HSIAO Roger, ZHANG Le, KARAKOS Damianos, MALLIDI Sri Harish, KARAFIÁT Martin, VESELÝ Karel, SZŐKE Igor, ZHANG Bing, NGUYEN Long and SCHWARTZ Richard. Progress in the BBN Keyword Search System for the DARPA RATS Program. In: Proceedings of Interspeech 2014. Singapore: International Speech Communication Association, 2014, pp. 959-963. ISBN 978-1-63439-435-2. Available from: http://www.isca-speech.org/archive/interspeech_2014/i14_0959.html
Czech title
Pokrok v BBN systému vyhledávání klíčových slov pro DARPA RATS program
Type
conference paper
Language
english
Authors
Ng Tim (Raytheon BBN)
Hsiao Roger (Raytheon BBN)
Zhang Le (Raytheon BBN)
Karakos Damianos (Raytheon BBN)
Mallidi Sri Harish (AmazonCom)
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Veselý Karel, Ing., Ph.D. (DCGM FIT BUT)
Szőke Igor, Ing., Ph.D. (DCGM FIT BUT)
Zhang Bing (Raytheon BBN)
Nguyen Long (Raytheon BBN)
Schwartz Richard (Raytheon BBN)
URL
Keywords

speech recognition, KWS, MLP, DNN

Abstract

This article is about the progress in the BBN Keyword Search System for the DARPA RATS Program (Robust Automatic Transcription of Speech).

Annotation

This paper presents a set of techniques that we used to improve our keyword search system for the third phase of the DARPA RATS (Robust Automatic Transcription of Speech) program, which seeks to advance state of the art detection capabilities on audio from highly degraded radio communication channels. The results for both Levantine and Farsi, which are the two target languages for the keyword search (KWS) task, are reported. About 13% absolute reduction in word error rate (from 70.2% to 57.6%) is achieved by using acoustic features derived from stacked Multi-Layer Perceptrons (MLP) and Deep Neural Network (DNN) acoustic models. In addition to score normalization and score/system combination for keyword search, we showed that the false alarm rate at the target false reject rate (15%) was reduced by about 1% (from 5.39% to 4.45%) by reducing the deletion errors of the speech-to-text system.

Published
2014
Pages
959-963
Proceedings
Proceedings of Interspeech 2014
Conference
Interspeech Conference, Singapur, SG
ISBN
978-1-63439-435-2
Publisher
International Speech Communication Association
Place
Singapore, SG
UT WoS
000395050100195
EID Scopus
BibTeX
@INPROCEEDINGS{FITPUB10744,
   author = "Tim Ng and Roger Hsiao and Le Zhang and Damianos Karakos and Harish Sri Mallidi and Martin Karafi\'{a}t and Karel Vesel\'{y} and Igor Sz\H{o}ke and Bing Zhang and Long Nguyen and Richard Schwartz",
   title = "Progress in the BBN Keyword Search System for the DARPA RATS Program",
   pages = "959--963",
   booktitle = "Proceedings of Interspeech 2014",
   year = 2014,
   location = "Singapore, SG",
   publisher = "International Speech Communication Association",
   ISBN = "978-1-63439-435-2",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/10744"
}
Back to top