Hang Su: Combining Speech and Speaker Recognition - A Joint Modeling Approach. University of California, Berkeley, USA 2018