cs.LGFeb 1, 2026

A Unified Matrix-Spectral Framework for Stability and Interpretability in Deep Learning

Authors: Ronald Katende

Organizations: Department of Mathematics, Kabale University, Kikungiri Hill, Katuna Road, 317, Kabale, Uganda

Abstract

We develop a unified matrix-spectral framework for analyzing stability and interpretability in deep neural networks. Representing networks as data-dependent products of linear operators reveals spectral quantities governing sensitivity to input perturbations, label noise, and training dynamics. We introduce a Global Matrix Stability Index that aggregates spectral information from Jacobians, parameter gradients, Neural Tangent Kernel operators, and loss Hessians into a single stability scale controlling forward sensitivity, attribution robustness, and optimization conditioning. We further show that spectral entropy refines classical operator-norm bounds by capturing typical, rather than purely worst-case, sensitivity. These quantities yield computable diagnostics and stability-oriented regularization principles. Synthetic experiments and controlled studies on MNIST, CIFAR-10, and CIFAR-100 confirm that modest spectral regularization substantially improves attribution stability even when global spectral summaries change little. The results establish a precise connection between spectral concentration and analytic stability, providing practical guidance for robustness-aware model design and training.

Figures & tables

Explore similar work

CardsList
  1. Non-normal spectral signatures of instability in neural network training dynamics

    May 22, 2026Souvik GhoshDynamical SystemsNeural Network Optimization

  2. Implicit Bias via Spectral Neural Networks and Non-linear Matrix Sensing

    Feb 27, 2024Hong T. M. Chu, Xiaoyu Dong, Subhro Ghosh +2Implicit RegularizationNeural Network Training Dynamics

  3. The Spectral Neuron

    Aug 8, 2026Alex ShtoffNeural Network RobustnessNeural Network Interpretability