Geometric Mean Metric Learning
arXiv:1607.05002
Abstract
We revisit the task of learning a Euclidean metric from data. We approach this problem from first principles and formulate it as a surprisingly simple optimization problem. Indeed, our formulation even admits a closed form solution. This solution possesses several very attractive properties: (i) an innate geometric appeal through the Riemannian geometry of positive definite matrices; (ii) ease of interpretability; and (iii) computational speed several orders of magnitude faster than the widely used LMNN and ITML methods. Furthermore, on standard benchmark datasets, our closed-form solution consistently attains higher classification accuracy.
7 pages, 4 figures
References in corpus (4)
Cited by in corpus (15)
- Low-rank geometric mean metric learning
- Learning to Approximate a Bregman Divergence
- Submodular Norms with Applications To Online Facility Location and Stochastic Probing
- Efficient Hierarchical Bayesian Inference for Spatio-temporal Regression Models in Neuroimaging
- Metric Learning from Imbalanced Data
- Adaptive Nearest Neighbor: A General Framework for Distance Metric Learning
- Multi-Facet Recommender Networks with Spherical Optimization
- Learning with symmetric positive definite matrices via generalized Bures-Wasserstein geometry
- A Riemannian Primal-dual Algorithm Based on Proximal Operator and its Application in Metric Learning
- Vector Transport Free Riemannian LBFGS for Optimization on Symmetric Positive Definite Matrix Manifolds
- Signed Graph Metric Learning via Gershgorin Disc Perfect Alignment
- Positive semidefinite support vector regression metric learning
- Isolation Kernel: The X Factor in Efficient and Effective Large Scale Online Kernel Learning
- A Probabilistic approach for Learning Embeddings without Supervision
- Isolation Distributional Kernel: A New Tool for Point & Group Anomaly Detection