Modeling Pronunciation Variation for Automatic Speech Recognition Rolduc, The Netherlands May 4-6, 1998

A Surficial Pronunciation Model

Eric Sven Ristad (1), Peter N. Yianilos (2)

(1) Mnemonic Technology, Inc.; (2) NEC Research Institute, Princeton, NJ, USA

We argue for a surficial pronunciation model: a model without underlying forms. The surficial model outperforms a traditional generative model by a significant margin on conversational speech (Switchboard) as well as on read speech (TIMIT). Our results suggest that the true mapping from underlying forms to surface forms is too complex to be accurately modeled using current techniques, and that we would be best served to model the surface forms directly.

Full Paper

Bibliographic reference.  Ristad, Eric Sven / Yianilos, Peter N. (1998): "A surficial pronunciation model", In MPV-1998, 117-120.