A Non-isotropic Probabilistic Take on Proxy-based Deep Metric Learning.
Michael KirchhofKarsten RothZeynep AkataEnkelejda KasneciPublished in: ECCV (26) (2022)
Keyphrases
- metric learning
- distance metric
- distance metric learning
- pairwise
- semi supervised
- dimensionality reduction
- learning tasks
- machine learning and pattern recognition
- person re identification
- nearest neighbor classification
- distance function
- mahalanobis metric
- multi task
- semi supervised clustering
- feature space
- probabilistic model
- maximum variance unfolding
- subject to linear constraints
- semi supervised learning
- euclidean distance
- low dimensional
- principal component analysis
- semi definite programming
- machine learning