Classifying high-dimensional Gaussian mixtures: Where kernel methods fail and neural networks succeed.
Maria RefinettiSebastian GoldtFlorent KrzakalaLenka ZdeborováPublished in: CoRR (2021)
Keyphrases
- kernel methods
- gaussian mixture
- high dimensional
- neural network
- kernel function
- feature space
- high dimensional feature space
- gaussian mixture model
- em algorithm
- probability density function
- closed form
- covariance matrix
- expectation maximization
- kernel principal component analysis
- kernel matrix
- support vector
- dimensionality reduction
- kernel pca
- support vector machine
- machine learning
- low dimensional
- pattern recognition
- similarity search
- nearest neighbor
- reproducing kernel hilbert space
- high dimensional data
- mixture model
- maximum likelihood
- input space
- data points
- training samples
- semi supervised
- density estimation
- probability distribution
- prior knowledge
- computer vision