UNSSOR: Unsupervised Neural Speech Separation by Leveraging Over-determined Training Mixtures.
Zhong-Qiu WangShinji WatanabePublished in: CoRR (2023)
Keyphrases
- supervised learning
- supervised training
- hearing impaired
- speech signal
- supervised methods
- neural network
- speech recognition
- training algorithm
- deep architectures
- blind separation
- test set
- unsupervised learning
- training examples
- training set
- training samples
- training process
- mixture model
- network architecture
- machine learning
- reinforcement learning
- artificial neural networks
- recurrent networks
- unsupervised methods
- language model
- heavy tailed
- unsupervised manner
- neural model
- automatic speech recognition
- bio inspired