Optimizing acoustic models for commercial speech recognition using foreground scores and data weighting.

Daniel Boies, Brian Strope, Mitchel Weintraub, Su-Lin Wu

Browse the full ICASSP paper archive.