Publication detail

Convolutive Bottleneck Network Features for LVCSR

VESELÝ, K. KARAFIÁT, M. GRÉZL, F.

Original Title

Convolutive Bottleneck Network Features for LVCSR

English Title

Convolutive Bottleneck Network Features for LVCSR

Type

conference paper

Language

en

Original Abstract

Workshop Article about novel features for tandem LVCSR system, which are based on Convolutive Bottleneck Network. It extends the previous work on Universal Context network by using linear bottleneck and expansion to Convolutive Bottleneck Network,so all the parameters are trained together.

English abstract

Workshop Article about novel features for tandem LVCSR system, which are based on Convolutive Bottleneck Network. It extends the previous work on Universal Context network by using linear bottleneck and expansion to Convolutive Bottleneck Network,so all the parameters are trained together.

Keywords

Bottleneck features, Tandem LVCSR system, linear bottleneck, Convolutional Bottleneck Network

RIV year

2011

Released

11.12.2011

Publisher

IEEE Signal Processing Society

Location

Big Island, Hawaii

ISBN

978-1-4673-0366-8

Book

Proceedings of ASRU 2011

Edition

NEUVEDEN

Edition number

NEUVEDEN

Pages from

42

Pages to

47

Pages count

6

URL

Documents

BibTex


@inproceedings{BUT76443,
  author="Karel {Veselý} and Martin {Karafiát} and František {Grézl}",
  title="Convolutive Bottleneck Network Features for LVCSR",
  annote="Workshop Article about novel features for tandem LVCSR system, which are based on
Convolutive Bottleneck Network. It extends the previous work on Universal Context
network by using linear bottleneck and expansion to Convolutive Bottleneck
Network,so all the parameters are trained together.",
  address="IEEE Signal Processing Society",
  booktitle="Proceedings of ASRU 2011",
  chapter="76443",
  edition="NEUVEDEN",
  howpublished="print",
  institution="IEEE Signal Processing Society",
  year="2011",
  month="december",
  pages="42--47",
  publisher="IEEE Signal Processing Society",
  type="conference paper"
}