Speaker Adaptation Based on PARAFAC2 of Transformation Matrices for Continuous Speech Recognition

Abstract

SUMMARYWe present an acoustic model adaptation method where the transformation matrix for a new speaker is given by the product of bases and a weight matrix. The bases are built from the parallel factor analysis 2 (PARAFAC2) of training speakers' transformation matrices. We perform continuous speech recognition experiments using the WSJ0 corpus. key words: maximum likelihood linear regression, parallel factor analysis, PARAFAC2, speaker adaptation, speech recognition