Domain Adaptation Using Suitable Pseudo Labels for Speech Enhancement and Dereverberation.
Lior FrenkelShlomo E. ChazanJacob GoldbergerPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2024)
Keyphrases
- domain adaptation
- speech enhancement
- noisy environments
- labeled data
- multiple sources
- noise reduction
- cross domain
- transfer learning
- training data
- semi supervised
- signal to noise ratio
- speech signal
- semi supervised learning
- single channel
- sentiment classification
- unlabeled data
- smoothing algorithm
- vocal tract
- test data
- pairwise
- document classification
- target domain
- similarity measure
- image coding
- class labels
- co training
- data mining
- linear prediction
- e government
- image segmentation