ISCA Archive SLAM 2013
ISCA Archive SLAM 2013

Slightly supervised adaptation of acoustic models on captioned BBC weather forecasts

Christian Mohr, Christian Saam, Kevin Kilgour, Jonas Gehring, Sebastian Stüker, Alex Waibel

In this paper we investigate the exploitation of loosely transcribed audio data, in the form of captions for weather forecast recordings, in order to adapt acoustic models for automatically transcribing these kinds of forecasts. We focus on dealing with inaccurate time stamps in the captions and the fact that they often deviate from the exact spoken word sequence in the forecasts. Furthermore, different adaptation algorithms are compared when incrementally increasing the amount of adaptation material, for example, by recording new forecasts on a daily basis.

Index Terms: speech recognition, acoustic model adaptation, slightly supervised training, loose transcripts, adaptation methods


Cite as: Mohr, C., Saam, C., Kilgour, K., Gehring, J., Stüker, S., Waibel, A. (2013) Slightly supervised adaptation of acoustic models on captioned BBC weather forecasts. Proc. First Workshop on Speech, Language and Audio in Multimedia (SLAM 2013), 32-36

@inproceedings{mohr13_slam,
  author={Christian Mohr and Christian Saam and Kevin Kilgour and Jonas Gehring and Sebastian Stüker and Alex Waibel},
  title={{Slightly supervised adaptation of acoustic models on captioned BBC weather forecasts}},
  year=2013,
  booktitle={Proc. First Workshop on Speech, Language and Audio in Multimedia (SLAM 2013)},
  pages={32--36}
}