About combining forward and backward-based decoders for selecting data for unsupervised training of acoustic models.

Denis Jouvet, Dominique Fohr

Browse the full Interspeech paper archive.