Abstract:Given a convex and differentiable objective $Q(\M)$ for a real symmetric matrix $\M$ in the positive definite (PD) cone -- used to compute Mahalanobis distances -- we propose a fast general metric learning framework that is entirely projection-free. We first assume that $\M$ resides in a space $\cS$ of generalized graph Laplacian matrices corresponding to balanced signed graphs. $\M \in \cS$ that is also PD is called a graph metric matrix. Unlike low-rank metric matrices common in the literature, $\cS$ includes the important diagonal-only matrices as a special case. The key theorem to circumvent full eigen-decomposition and enable fast metric matrix optimization is Gershgorin disc perfect alignment (GDPA): given $\M \in \cS$ and diagonal matrix $§$, where $S_{ii} = 1/v_i$ and $\v$ is $\M$'s first eigenvector, we prove that Gershgorin disc left-ends of similarity transform $\B = §\M §^{-1}$ are perfectly aligned at the smallest eigenvalue $\lambda_{\min}$. Using this theorem, we replace the PD cone constraint in the metric learning problem with tightest possible linear constraints per iteration, so that the alternating optimization of the diagonal / off-diagonal terms in $\M$ can be solved efficiently as linear programs via the Frank-Wolfe method. We update $\v$ using Locally Optimal Block Preconditioned Conjugate Gradient (LOBPCG) with warm start as entries in $\M$ are optimized successively. Experiments show that our graph metric optimization is significantly faster than cone-projection schemes, and produces competitive binary classification performance.

The Cheeger cut and Cheeger problem in metric graphs

Cheeger's Cut, Maxcut and the Spectral Theory Of1-Laplacian on Graphs

An overview on the Cheeger problem

Mean Distance on Metric Graphs

A Cheeger inequality for the lower spectral gap

Towards a Calculus for Non-Linear Spectral Gaps [Extended Abstract]

The Cheeger problem in abstract measure spaces

On the Cheeger constant for distance-regular graphs

Metric Graph Approximations of Geodesic Spaces

Cheeger cuts and Robin spectral minimal partitions of metric graphs

Graph minors and metric spaces

Notes on the Cheeger and Colding version of the Reifenberg theorem for metric spaces

Signed Graph Metric Learning via Gershgorin Disc Perfect Alignment

Graph Metric Learning Via Gershgorin Disc Alignment

On the Cheeger sets in strips and non-convex domains

On the Houdré-Tetali conjecture about an isoperimetric constant of graphs

The Cheeger constant of a quantum graph

Graph GOSPA Metric: A Metric to Measure the Discrepancy Between Graphs of Different Sizes

-Metric Graphs: Radius, Diameter and all Eccentricities

Metric for comparing graphs with ordered vertices based on the maximum common subgraph