A denoised embedding space of genetic perturbation using Deep Metric Learning.
Minjae JuSanghoon LeeJaewoo KangPublished in: BigComp (2022)
Keyphrases
- metric learning
- embedding space
- dimensionality reduction
- kernel matrix
- manifold structure
- manifold learning
- semi supervised
- low dimensional
- distance metric
- graph embedding
- input space
- denoising
- feature space
- data points
- high dimensional
- euclidean space
- high dimensional data
- principal component analysis
- nonlinear dimensionality reduction
- pairwise
- distance function
- euclidean distance
- pattern recognition
- multi task
- positive semidefinite
- semidefinite programming
- learning tasks
- data representation
- unsupervised learning
- feature extraction
- linear discriminant analysis
- locally linear embedding
- semi supervised learning
- geometric structure
- geodesic distance
- pairwise constraints
- dimensionality reduction methods
- sparse representation
- nearest neighbor
- feature selection
- similarity search
- unlabeled data
- active learning
- decision trees
- image processing