UniSpeech: Unified Speech Representation Learning
              with Labeled and Unlabeled Data

   Chengyi Wang * 1 Yu Wu 2 Yao Qian 2 Kenichi Kumatani 2 Shujie Liu 2 Furu Wei 2 Michael Zeng 2
                      Xuedong Huang 2

            Abstract                majority of the nearly 7000 languages spoken worldwide
  In this paper, we propose a unified pre-training      (Katzner & Miller, 2002). This poses a real challenge for
  approach called UniSpeech to learn speech repre-      buildi ...