Speaker Adaptation Based on Nonlinear Spectral Transform for Speech Recognition

This paper proposes a speaker adaptation technique using a nonlinear spectral transform based on GMMs. One of the most popular forms of speaker adaptation is based on linear transforms, e.g., MLLR. Although MLLR uses multiple transforms according to regression classes, only a single linear transform is applied to each state. The proposed method performs nonlinear speaker adaptation based on a new likelihood function combining HMMs for recognition with GMMs for spectral transform. Moreover, the dependency of transforms on context can also be estimated in an integrated ML fashion. The proposed technique outperformed conventional approaches in phoneme-recognition experiments