Efficient speech recognition using subvector quantization and discrete-mixture HMMs | IEEE Conference Publication | IEEE Xplore