CONF sivadas-rr-03-57p/IDIAP On Use of Task Independent Training Data in Tandem Feature Extraction Sivadas, Sunil Hermansky, Hynek EXTERNAL http://publications.idiap.ch/attachments/reports/2003/rr-03-57.pdf PUBLIC http://publications.idiap.ch/index.php/publications/showcite/sivadas-rr-03-57 Related documents Proceedings of the 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP-04) 2004 Montreal, Canada May 2004 The problem we address in this paper is, whether the feature extraction module trained on large amounts of task independent data, can improve the performance of stochastic models? We show that when there is only a small amount of task specific training data available, tandem features trained on task independent data give considerable improvement over Perceptual Linear Prediction (PLP) cepstral features in Hidden Markov Model (HMM) based speech recognition systems. REPORT sivadas-rr-03-57/IDIAP On Use of Task Independent Training Data in Tandem Feature Extraction Sivadas, Sunil Hermansky, Hynek EXTERNAL http://publications.idiap.ch/attachments/reports/2003/rr-03-57.pdf PUBLIC Idiap-RR-57-2003 2003 IDIAP Martigny, Switzerland The problem we address in this paper is, whether the feature extraction module trained on large amounts of task independent data, can improve the performance of stochastic models? We show that when there is only a small amount of task specific training data available, tandem features trained on task independent data give considerable improvement over Perceptual Linear Prediction (PLP) cepstral features in Hidden Markov Model (HMM) based speech recognition systems.