Lattice-based discriminative adaptation integrated into LF-MMI enables robust unsupervised test-time adaptation of neural acoustic models on mismatched speech data.
In feature-space adaptation , trans- formations of acoustic features are estimated to maximise t he log-likelihood of the adaptation data [1, 2]
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Lattice-Based Unsupervised Test-Time Adaptation of Neural Network Acoustic Models
Lattice-based discriminative adaptation integrated into LF-MMI enables robust unsupervised test-time adaptation of neural acoustic models on mismatched speech data.