KL-Divergence Regularized Deep Neural Network Adaptation for Low-Resource Speaker-Dependent Speech Enhancement.
Li ChaiJun DuChin-Hui LeePublished in: INTERSPEECH (2019)
Keyphrases
- kl divergence
- neural network
- kullback leibler divergence
- speaker adaptation
- mahalanobis distance
- information theoretic
- exponential family
- noisy environments
- pattern recognition
- vocal tract
- speech signal
- gaussian distribution
- speech recognition
- maximum likelihood
- speaker identification
- noise reduction
- denoising
- least squares
- closed form
- mutual information