Unsupervised learning of acoustic features via deep canonical correlation analysis.
Weiran WangRaman AroraKaren LivescuJeff A. BilmesPublished in: ICASSP (2015)
Keyphrases
- canonical correlation analysis
- acoustic features
- unsupervised learning
- speaker verification
- music information retrieval
- automatic speech recognition
- visual features
- least squares
- speech signal
- supervised learning
- dimensionality reduction
- partial least squares
- dimensionality reduction methods
- cross correlation
- semi supervised
- audio features
- dimension reduction
- speech recognition
- feature selection
- information retrieval
- high dimensional
- generative model
- expectation maximization
- topic models
- information retrieval systems
- classification accuracy
- object recognition