Unsupervised Pre-Training of Bidirectional Speech Encoders via Masked Reconstruction.
Weiran WangQingming TangKaren LivescuPublished in: ICASSP (2020)
Keyphrases
- supervised learning
- hearing impaired
- supervised training
- unsupervised learning
- training phase
- deep architectures
- text to speech
- semi supervised
- training process
- speech recognition
- test set
- training set
- three dimensional
- machine learning
- maximum likelihood
- human computer interaction
- neural network
- online learning
- object detection
- video compression
- speech signal
- high resolution
- reconstruction method
- compressive sensing
- unsupervised manner
- speech synthesis
- support vector