Dong Yu, Shizhen Wang, Jinyu Li, and Li Deng
It is widely known that the quality of confidence measure is critical for speech applications. In this paper, we present our recent work on improving word confidence scores by calibrating them using a small set of calibration data when only the recognized word sequence and associated raw confidence scores are made available. The core of our technique is the maximum entropy model with distribution constraints which naturally and effectively make use of the word distribution, the raw confidence-score distribution, and the context information. We demonstrate the effectiveness of our approach by showing that it can achieve relative 38% mean square error (MSE), 39% negative normalized likelihood (NNLL), and 23% equal error rate (EER) reduction on a voice mail transcription data set and relative 35% MSE, 45% NNLL, and 35% EER reduction on a command and control data set.
© 2008 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE. http://www.ieee.org/