Mixture Factorized Auto-Encoder for Unsupervised Hierarchical Deep Factorization of Speech Signal.
Zhiyuan PengSiyuan FengTan LeePublished in: ICASSP (2020)
Keyphrases
- speech signal
- speech recognition
- matrix factorization
- automatic speech recognition
- unsupervised learning
- noisy environments
- background noise
- non stationary
- linear prediction
- speaker identification
- hidden markov models
- vocal tract
- expectation maximization
- speaker recognition
- spectral analysis
- mixture model
- deep belief networks
- automatic speech recognition systems
- bit rate
- gaussian mixture model
- acoustic features
- measurement matrix
- speech enhancement
- information retrieval