Purely sequence-trained neural networks for ASR based on lattice-free MMI

INTERSPEECH 2016 2016 Daniel PoveyVijayaditya PeddintiDaniel GalvezPegah GhahrmaniVimal ManoharXingyu NaYiming WangSanjeev Khudanpur

In this paper we describe a method to perform sequence-discriminative training of neural network acoustic models without the need for frame-level cross-entropy pre-training. We use the lattice-free version of the maximum mutual information (MMI) criterion: LF-MMI... (read more)

PDF

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Speech Recognition WSJ eval92 tdnn + chain Percentage error 2.32 # 1