The performance of keyword search systems depends heavily on the quality of confidence scores. In this work, a novel discriminative score calibration method has been proposed. By training an MLP classifier employing the word posterior probability and several novel normalized scores, we can obtain a relative improvement of 4.67% for the actual term-weighted value (ATWV) metric on the OpenKWS15 development test dataset. In addition, a LSTM-CTC based keyword verification method has been proposed to supply extra acoustic information. After the information is added, a further improvement of 7.05% over the baseline can be observed.
Cite as: Lv, Z., Cai, M., Zhang, W.-Q., Liu, J. (2016) A Novel Discriminative Score Calibration Method for Keyword Search. Proc. Interspeech 2016, 745-749, doi: 10.21437/Interspeech.2016-606
@inproceedings{lv16_interspeech,
author={Zhiqiang Lv and Meng Cai and Wei-Qiang Zhang and Jia Liu},
title={{A Novel Discriminative Score Calibration Method for Keyword Search}},
year=2016,
booktitle={Proc. Interspeech 2016},
pages={745--749},
doi={10.21437/Interspeech.2016-606}
}