Iterative pseudo-forced alignment by acoustic CTC loss for self-supervised ASR domain adaptation.
Fernando LópezJordi LuquePublished in: CoRR (2022)
Keyphrases
- domain adaptation
- manifold alignment
- cross domain
- transfer learning
- multiple sources
- labeled data
- sentiment classification
- semi supervised learning
- automatic speech recognition
- semi supervised
- speech recognition
- covariate shift
- document classification
- test data
- databases
- text classification
- unlabeled data
- data sets
- general purpose
- pairwise
- training set
- speech signal
- target domain
- co training
- data analysis
- information retrieval
- machine learning
- data mining