Transfer Learning for Robust Low-Resource Children's Speech ASR with Transformers and Source-Filter Warping.
Jenthe ThienpondtKris DemuynckPublished in: CoRR (2022)
Keyphrases
- transfer learning
- automatic speech recognition
- target domain
- speech recognition
- noisy environments
- learning tasks
- knowledge transfer
- source domain
- domain adaptation
- cross domain
- labeled data
- active learning
- reinforcement learning
- multi task learning
- hearing impaired
- multiple sources
- previously learned
- multi task
- machine learning
- noise reduction
- machine learning algorithms
- semi supervised learning
- manifold alignment
- speech signal
- text mining
- learning algorithm
- transferring knowledge
- text classification
- semi supervised
- data points
- spontaneous speech
- information retrieval