ISOMAP-KL: a parametric approach for unsupervised metric learning.
Alaor Cervati NetoAlexandre L. M. LevadaPublished in: SIBGRAPI (2020)
Keyphrases
- metric learning
- semi supervised
- maximum variance unfolding
- dimensionality reduction
- distance metric
- manifold learning
- distance metric learning
- nonlinear dimensionality reduction
- unsupervised learning
- fully supervised
- pairwise
- learning tasks
- euclidean distance
- kernel matrix
- semi supervised learning
- distance function
- mahalanobis distance
- supervised learning
- dimension reduction
- data visualization
- locally linear embedding
- feature space
- machine learning
- laplacian eigenmaps
- unlabeled data
- multi task
- low dimensional
- distance measure
- pattern recognition
- feature mapping
- labeled data
- high dimensional
- image processing
- high dimensional data
- text mining
- principal component analysis
- active learning
- data sets
- mahalanobis metric
- subject to linear constraints