Unsupervised Ground Metric Learning Using Wasserstein Singular Vectors.
Geert-Jan HuizingLaura CantiniGabriel PeyréPublished in: ICML (2022)
Keyphrases
- metric learning
- semi supervised
- singular vectors
- singular value decomposition
- distance metric
- dimensionality reduction
- pairwise
- learning tasks
- semi supervised learning
- singular values
- supervised learning
- unsupervised learning
- distance function
- multi task
- active learning
- euclidean distance
- low rank
- kernel matrix
- feature space
- machine learning
- neural network
- least squares
- pattern recognition
- decision trees