Speaker adaptation based on the multilinear decomposition of training speaker models.

Yongwon Jeong

Browse the full ICASSP paper archive.