Distance-Ratio-Based Formulation for Metric Learning.
Hyeongji KimPekka ParviainenKetil MaldePublished in: CoRR (2022)
Keyphrases
- metric learning
- distance metric
- distance function
- mahalanobis distance
- distance metric learning
- euclidean distance
- machine learning and pattern recognition
- kernel matrix
- dimensionality reduction
- pairwise
- feature space
- semi supervised
- distance measure
- pairwise similarities
- knn
- multi task
- nearest neighbor
- data points
- kernel learning
- learning tasks
- feature extraction
- semi definite programming
- data mining
- prior knowledge
- data sets
- similarity search
- high dimensional data
- mutual information
- feature vectors