Ninth International Conference on Spoken Language Processing

Pittsburgh, PA, USA
September 17-21, 2006

Improving Speech Recognition Accuracy with Multi-Confidence Thresholding

Shuangyu Chang

Tellme Networks Inc., USA

Confidence-based thresholding plays an important role in practical speech recognition applications. Most previous works have focused on directly improving confidence estimation within the recognition engine. A complementary approach that does not require access to recognizer internal is to optimize confidence threshold settings. This paper describes a general multi-confidence thresholding algorithm that automatically learns different confidence thresholds for different utterances, based on discreet or continuous features associated with a speech utterance. The algorithm can be applied to any speech recognition engine with a confidence output. A learned multi-threshold setting is guaranteed to perform at least as well as a baseline singlethreshold system on training data. A significant improvement on overall accuracy can often be obtained on test data, as demonstrated with experiments on two real-world applications.

Full Paper

Bibliographic reference.  Chang, Shuangyu (2006): "Improving speech recognition accuracy with multi-confidence thresholding", In INTERSPEECH-2006, paper 1346-Wed1CaP.3.