GMML
Geometric Mean Metric Learning
view repo
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.
READ FULL TEXT
Over the past few years, symmetric positive definite (SPD) matrices have...
read it
Data encoded as symmetric positive definite (SPD) matrices frequently ar...
read it
This paper proposes an inexpensive way to learn an effective dissimilari...
read it
To achieve a low computational cost when performing online metric learni...
read it
In this paper, we present new results on the Riemannian geometry of symm...
read it
For decades, it has been widely accepted that the gold standard for two-...
read it
Metric space magnitude, an active subject of research in algebraic topol...
read it
Geometric Mean Metric Learning
Comments
There are no comments yet.