Rethinking preventing class-collapsing in metric learning with margin-based losses.
Elad LeviTete XiaoXiaolong WangTrevor DarrellPublished in: ICCV (2021)
Keyphrases
- metric learning
- distance metric
- distance metric learning
- pairwise constraints
- learning tasks
- pairwise
- semi supervised
- machine learning and pattern recognition
- multi task
- dimensionality reduction
- distance function
- multi class
- mahalanobis metric
- feature space
- semantic information
- class specific
- neural network
- semi supervised learning
- labeled data
- pattern recognition