Self-distillation with Online Diffusion on Batch Manifolds Improves Deep Metric Learning.
Zelong ZengFan YangHong LiuShin'ichi SatohPublished in: CoRR (2022)
Keyphrases
- metric learning
- locally linear
- distance metric
- distance metric learning
- manifold structure
- semi supervised
- feature space
- pairwise
- machine learning and pattern recognition
- multi task
- dimensionality reduction
- distance function
- manifold learning
- learning tasks
- kernel matrix
- mahalanobis metric
- subject to linear constraints
- machine learning
- semi supervised learning
- high dimensional data
- low dimensional
- feature extraction