Beyond cross-entropy: towards better frame-level objective functions for deep neural network training in automatic speech recognition.

Zhen Huang, Jinyu Li, Chao Weng, Chin-Hui Lee

Browse the full Interspeech paper archive.