Using untranscribed user utterances for improving language models based on confidence scoring.

Mikio Nakano, Timothy J. Hazen

Browse the full Interspeech paper archive.