Feature Dimensionality Reduction with Variational Autoencoders in Deep Bayesian Active Learning.
Pinar Ezgi ÇölSeyda ErtekinPublished in: SIU (2021)
Keyphrases
- dimensionality reduction
- active learning
- preprocessing step
- principal component analysis
- feature extraction
- pattern recognition
- high dimensionality
- denoising
- semi supervised
- bayesian networks
- data points
- high dimensional data
- low dimensional
- pattern recognition and machine learning
- belief nets
- batch mode
- unsupervised learning
- labeled data
- semi supervised learning
- data sets
- learning strategies
- supervised learning
- random sampling
- high dimensional
- bayesian learning
- kernel learning
- training set
- structure preserving
- image processing
- active learning strategies
- learning algorithm