Divide and Conquer the Embedding Space for Metric Learning.
Artsiom SanakoyeuVadim TschernezkiUta BüchlerBjörn OmmerPublished in: CVPR (2019)
Keyphrases
- metric learning
- embedding space
- dimensionality reduction
- kernel matrix
- manifold structure
- manifold learning
- distance metric
- low dimensional
- graph embedding
- data points
- semi supervised
- high dimensional
- euclidean space
- feature space
- input space
- principal component analysis
- euclidean distance
- geodesic distance
- pairwise
- high dimensional data
- linear discriminant analysis
- distance function
- unsupervised learning
- pattern recognition
- learning tasks
- nonlinear dimensionality reduction
- dimensionality reduction methods
- geometric structure
- semidefinite programming
- feature extraction
- positive semidefinite
- multi task
- pairwise constraints
- data representation
- locally linear embedding
- dimension reduction
- semi supervised learning
- multi class
- feature selection
- neural network