Smooth Proxy-Anchor Loss for Noisy Metric Learning.
Carlos RoigDavid VarasIssey MasudaJuan Carlos RiveiroElisenda Bou-BalustPublished in: CoRR (2020)
Keyphrases
- metric learning
- distance metric learning
- distance metric
- learning tasks
- dimensionality reduction
- machine learning and pattern recognition
- semi supervised
- pairwise
- distance function
- person re identification
- multi task
- nearest neighbor classification
- feature space
- semi supervised clustering
- semi supervised learning
- learning algorithm
- domain knowledge
- semi definite programming
- mahalanobis metric
- labeled data
- data analysis
- pattern recognition
- feature extraction
- knowledge base
- neural network
- maximum variance unfolding