Posterior Refinement on Metric Matrix Improves Generalization Bound in Metric Learning.
Mingda WangCanqian YangYi XuPublished in: ECCV (26) (2022)
Keyphrases
- metric learning
- positive semidefinite
- positive semi definite
- distance metric
- distance metric learning
- semi supervised
- learning tasks
- distance function
- pairwise
- nearest neighbor classification
- kernel matrix
- mahalanobis metric
- multi task
- dimensionality reduction
- subject to linear constraints
- learned metric
- mahalanobis distance
- semi supervised learning
- feature space
- semi definite programming
- sparse metric learning
- gaussian process
- low rank
- covariance matrix
- labeled data
- active learning
- face recognition
- neural network