Improving Acoustic Models by Watching Television
Michael Witbrock, Alexander G. Hauptmann · 1998
Obtaining sufficient labelled training data is a persistent difficulty for speech recognition research. Although well transcribed data is expensive to produce, there is a constant stream of challenging speech data and poor transcription broadcast as closed-captioned television. We describe a reliable unsupervised method for identifying accurately transcribed sections of these broadcasts, and show how these segments can be used to train a recognition system. Starting from acoustic models trained on the Wall Street Journal database, a single iteration of our training method reduced the word error rate on an independent broadcast television news test set from 62.2 % to 59.5%. This paper is based on work supported by the National Science Foundation, DARPA and NASA under NSF Cooperative agreement No. IRI-9411299. We thank Justsystem Corporation for supporting the preparation of the paper. Introduction Current speech recognition research is characterized by its reliance on data. The stati...