Incorporating mask modelling for noise-robust automatic speech recognition | IEEE Conference Publication | IEEE Xplore