Direct adaptation of hybrid DNN/HMM model for fast speaker adaptation in LVCSR based on speaker code | IEEE Conference Publication | IEEE Xplore