A Gaussian-kernel diffusion operator on feature clouds yields closed-form class affinities and spectra in Gaussian models, with provably smooth observables under perturbations.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
other 1
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2representative citing papers
A condition-number principle shows that small suboptimality in admissible prototype clustering objectives implies small misclassification error when the condition number is low, with phase transitions for exact recovery.
citing papers explorer
-
Diffusion Operator Geometry of Feedforward Representations
A Gaussian-kernel diffusion operator on feature clouds yields closed-form class affinities and spectra in Gaussian models, with provably smooth observables under perturbations.
-
The Condition-Number Principle for Prototype Clustering
A condition-number principle shows that small suboptimality in admissible prototype clustering objectives implies small misclassification error when the condition number is low, with phase transitions for exact recovery.