PCAAE: Principal Component Analysis Autoencoder for organising the latent space of generative networks.
Chi-Hieu PhamSaïd LadjalAlasdair NewsonPublished in: CoRR (2020)
Keyphrases
- latent space
- principal component analysis
- low dimensional
- lower dimensional
- generative model
- gaussian process latent variable models
- dimensionality reduction
- latent variables
- high dimensional
- covariance matrix
- manifold learning
- linear discriminant analysis
- face recognition
- dimension reduction
- parameter space
- gaussian process
- feature space
- unsupervised learning
- prior knowledge
- data sets
- singular value decomposition
- gaussian mixture
- negative matrix factorization
- random projections
- machine learning
- model selection
- computer vision
- higher dimensional
- pairwise
- sample size