Tied-state based discriminative training of context-expanded region-dependent feature transforms for LVCSR.

Zhi-Jie Yan, Qiang Huo, Jian Xu, Yu Zhang

Browse the full ICASSP paper archive.