CONF stephenson02c/IDIAP Dynamic Bayesian Network Based Speech Recognition with Pitch and Energy as Auxiliary Variables Stephenson, Todd Andrew Escofet, Jaume Magimai-Doss, Mathew Bourlard, Hervé EXTERNAL http://publications.idiap.ch/attachments/papers/2002/todd-nnsp2002.pdf PUBLIC http://publications.idiap.ch/index.php/publications/showcite/stephenson02b Related documents 2002 IEEE International Workshop on Neural Networks for for Signal Processing (NNSP~2002) 2002 Martigny, Switzerland September 2002 637-646 Pitch and energy are two fundamental features describing speech, having importance in human speech recognition. However, when incorporated as features in automatic speech recognition (ASR,',','), they usually result in a significant degradation on recognition performance due to the noise inherent in estimating or modeling them. In this paper, we show experimentally how this can be corrected by either conditioning the emission distributions upon these features or by marginalizing out these features in recognition. Since this is not obvious to do with standard hidden Markov models (HMMs,',','), this work has been performed in the framework of dynamic Bayesian networks (DBNs,',','), resulting in more flexibility in defining the topology of the emission distributions and in specifying whether variables should be marginalized out. REPORT stephenson02b/IDIAP Dynamic Bayesian Network Based Speech Recognition with Pitch and Energy as Auxiliary Variables Stephenson, Todd Andrew Escofet, Jaume Magimai-Doss, Mathew Bourlard, Hervé EXTERNAL http://publications.idiap.ch/attachments/reports/2002/rr02-24.pdf PUBLIC Idiap-RR-24-2002 2002 IDIAP In ``2002 IEEE International Workshop on Neural Networks for Signal Processing (NNSP~2002)'', 2002 Pitch and energy are two fundamental features describing speech, having importance in human speech recognition. However, when incorporated as features in automatic speech recognition (ASR,',','), they usually result in a significant degradation on recognition performance due to the noise inherent in estimating or modeling them. In this paper, we show experimentally how this can be corrected by either conditioning the emission distributions upon these features or by marginalizing out these features in recognition. Since this is not obvious to do with standard hidden Markov models (HMMs,',','), this work has been performed in the framework of dynamic Bayesian networks (DBNs,',','), resulting in more flexibility in defining the topology of the emission distributions and in specifying whether variables should be marginalized out.