A comparison of self-supervised speech representations as input features for unsupervised acoustic word embeddings.
Lisa van StadenHerman KamperPublished in: CoRR (2020)
Keyphrases
- input features
- speech recognition systems
- speech recognition
- speech sounds
- machine learning algorithms
- classification algorithm
- support vector machine
- co occurrence
- feature selection
- speech signal
- activation function
- feature values
- class labels
- feature subset
- automatic speech recognition
- semi supervised
- text mining
- supervised learning