wav2vec 2.0: A Framework for Self-Supervised
       Learning of Speech Representations

     Alexei Baevski    Henry Zhou     Abdelrahman Mohamed        Michael Auli
            {abaevski,henryzhou7,abdo,michaelauli}@fb.com
                       Facebook AI

                       Abstract
      We show for the first time that learning powerful representations from speech
      audio alone followed by fine-tuning on transcribed speech can outperform the best
      semi-supervised methods while  ...