Articulatory and Stacked Bottleneck Features for Low Resource Speech Recognition

doi 10.21437/interspeech.2018-2226