MathematicsFeatured paper
Spectral Methods for Sparse Network Recovery with Noisy Observations
Abstract
The paper develops a spectral recovery framework for sparse networks observed with structured noise. By combining thresholded eigenvector estimation with a stability correction step, the method improves edge recovery in settings where standard convex relaxations become unstable. The results are relevant for applied settings that rely on recovering latent graph structure from incomplete or contaminated measurements.
Publication Metadata
Keywords
spectral methodsnetwork recoverygraph inferencenoisy observations
Citation
Hannah Klein; Owen Bastien (2026). Spectral Methods for Sparse Network Recovery with Noisy Observations. Annals of Computational Mathematics. 10.48211/insight.math.2026.4108
Related Papers