Sequence generation error (SGE) minimization based deep neural networks training for text-to-speech synthesis.

Yuchen Fan, Yao Qian, Frank K. Soong, Lei He

Browse the full Interspeech paper archive.