Data selection from multiple ASR systems' hypotheses for unsupervised acoustic model training.

Sheng Li, Yuya Akita, Tatsuya Kawahara

Browse the full ICASSP paper archive.