Regularized linear autoencoders recover the principal components, eventually.
Xuchan BaoJames LucasSushant SachdevaRoger B. GrossePublished in: NeurIPS (2020)
Keyphrases
- principal components
- principal component analysis
- denoising
- multivariate data
- hyperplane
- kernel space
- principal components analysis
- correlation coefficient
- principal component regression
- dimensionality reduction
- spectral data
- feature set
- handwritten digits
- independent components
- kernel principal component analysis
- least squares
- decision trees
- feature selection
- data sets