Skip to main content

Table 1 Notations used in the article

From: Speaker adaptation in the maximum a posteriori framework based on the probabilistic 2-mode analysis of training models

Notation

Meaning

r

Index for the mixture component (1,…,R)

s

Index for the training speaker (1,…,S)

D

Dimension of the acoustic feature vector

μ

HMM mean vector

M

Matrix representation of HMM mean vector

ℳ

Tensor representation of training models

G

Core tensor

U

Mode matrix, factor loading matrix

w

Weight vector

w

Weight matrix, latent matrix

C, Ψ, Σ

covariance matrix

E

Error matrix