We apply this method to train a pronunciation model for recognition on conversational speech, resulting in signifi- cant improvements in recognition performance ...

This technique allows us to effectively estimate parameters of a factor WFST using relatively small amounts of data, if the factor is small. Our approach ...

One of the most popular speech recognition architectures con- sists of multiple components (like the acoustic, pronunciation and language models) that are ...

We apply this method to train a pronunciation model for recognition on conversational speech, resulting in signifi-cant improvements in recognition performance ...

Discriminative Training of WFST Factors with Application to Pronunciation Modeling. Jyothi, P., Fosler-Lussier, E., & Livescu, K. In Proceedings of ...

Discriminative training of WFST factors with application to pronunciation modeling. P. Jyothi, E. Fosler-Lussier, and K. Livescu. INTERSPEECH, page 1961 ...

The recognizer is represented as WFST factors or components, and they learn the parameters of the arcs from the pronunciation lexicon WFST in isolation, or ...

This paper describes an approach to efficiently construct, and discriminatively train, a weighted finite state transducer. (WFST) representation for an ...

2013. TLDR. A discriminative training approach is proposed that allows selective training of WFST factors within an ASR system based on WF STs, and a ...

PDF | This paper describes an approach to efficiently construct, and discriminatively train, a weighted finite state transducer (WFST) representation.