Exploiting long-range temporal dynamics of speech for noise-robust speaker recognition | IEEE Conference Publication | IEEE Xplore