Unsupervised adaptation is the most convenient mode for the user of a speech recognition system. However the performance of unsupervised adaptation is worse than that of the supervised mode because of the recognition errors. This paper introduces a kind of word-lattice based confidence measure to evaluate the reliability of the recognition result and discard the uncertain parts from the adaptation speech. Experiments demonstrate that the confidence can improve the performance of unsupervised adaptation considerably.
Cite as: Li, H., Liu, J., Liu, R. (2000) Confidence measure based unsupervised speaker adaptation. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 2, 819-822, doi: 10.21437/ICSLP.2000-395
@inproceedings{li00f_icslp,
author={Husheng Li and Jia Liu and Runsheng Liu},
title={{Confidence measure based unsupervised speaker adaptation}},
year=2000,
booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
pages={vol. 2, 819-822},
doi={10.21437/ICSLP.2000-395}
}