5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

How Effective Is Unsupervised Data Collection For Children's Speech Recognition?

Gregory Aist (1), Peggy Chan (1), Xuedong Huang (2), Li Jiang (2), Rebecca Kennedy (1), DeWitt Latimer (1), Jack Mostow (1), Calvin Yeung (1)

(1) Carnegie Mellon University, USA
(2) Microsoft Research, USA

Children present a unique challenge to automatic speech recognition. Today's state-of-the-art speech recognition systems still have problems handling children's speech because acoustic models are trained on data collected from adult speech. In this paper we describe an inexpensive way to mend this problem. We collected children's speech when they interact with an automated reading tutor. These data are subsequently transcribed by a speech recognition system and automatically filtered. We studied how to use these automatically collected data to improve children's speech recognition system's performance. Experiments indicate that automatically collected data can reduce the error rate significantly on children's speech.

Full Paper

Bibliographic reference.  Aist, Gregory / Chan, Peggy / Huang, Xuedong / Jiang, Li / Kennedy, Rebecca / Latimer, DeWitt / Mostow, Jack / Yeung, Calvin (1998): "How effective is unsupervised data collection for children's speech recognition?", In ICSLP-1998, paper 0929.