Transfer learning through perturbation-based in-domain spectrogram augmentation for adult speech recognition.
Virender KadyanPuneet BawaPublished in: Neural Comput. Appl. (2022)
Keyphrases
- transfer learning
- speech recognition
- speech signal
- automatic speech recognition
- cross domain
- knowledge transfer
- hidden markov models
- domain adaptation
- target domain
- reinforcement learning
- machine learning
- language model
- labeled data
- speech synthesis
- text classification
- noisy environments
- collaborative filtering
- speech recognizer
- text categorization
- transfer knowledge
- speech recognition technology
- speaker identification
- active learning
- speaker independent
- pattern recognition
- speech recognition systems
- cross lingual
- semi supervised learning
- computer vision
- learning algorithm
- text mining
- data analysis
- training data
- data sets
- machine learning algorithms
- speaker adaptation
- artificial intelligence
- prior knowledge
- domain specific
- error rate