ITRW on Adaptation Methods for Speech Recognition August 29-30, 2001 Sophia Antipolis, France

Unsupervised Language and Acoustic Model Adaptation for Cross Domain Portability

Diego Giuliani and Marcello Federico

ITC-irst - Centro per la Ricerca Scientifica e Tecnologica, Povo, Trento, Italy

This work investigates the task of porting a broadcast news recognition system to a conversational speech domain, for which only untranscribed acoustic data are available. An iterative adaptation procedure is proposed that alternatively generates automatic speech transcriptions and performs acoustic and language model adaptation. The procedure was applied on a tourist-information conversational domain, for which 8 hours of audio data were available for development and 2 hours for testing. On the test set, the broadcast news system yields a word-error-rate of 51.0% while a task specific system achieves a word-error-rate of 21.2%. Unsupervised porting experiments allowed to reduce the gap between the two reference systems by 61%.

Full Paper

Bibliographic reference.  Giuliani, Diego / Federico, Marcello (2001): "Unsupervised language and acoustic model adaptation for cross domain portability", In Adaptation-2001, 183-186.