Contractivity of neural ODEs: an eigenvalue optimization problem

TL;DR

This paper introduces a new methodology for analyzing the contractivity of neural ODEs through eigenvalue optimization, enabling stability assessment and control in neural network models.

Contribution

It develops a two-level nested approach to solve a key eigenvalue optimization problem related to neural ODE contractivity, extending to multilayer and time-dependent cases.

Findings

Effective eigenvalue optimization method for neural ODE stability

Application to stabilizing neural ODEs in MNIST classification

Extension to multilayer and time-dependent neural ODEs

Abstract

We propose a novel methodology to solve a key eigenvalue optimization problem which arises in the contractivity analysis of neural ODEs. When looking at contractivity properties of a one layer weight-tied neural ODE $\dot{u}(t)=\sigma(Au(t)+b)$ (with $u,b \in {\mathbb R}^n$, $A$ is a given $n \times n$ matrix, $\sigma : {\mathbb R} \to {\mathbb R}$ denotes an activation function and for a vector $z \in {\mathbb R}^n$, $\sigma(z) \in {\mathbb R}^n$ has to be interpreted entry-wise), we are led to study the logarithmic norm of a set of products of type $D A$, where $D$ is a diagonal matrix such that ${\mathrm{diag}}(D) \in \sigma'({\mathbb R}^n)$. Specifically, given a real number $c$ (usually $c=0$), the problem consists in finding the largest positive interval $\text{I}\subseteq \mathbb [0,\infty)$ such that the logarithmic norm $\mu(DA) \le c$ for all diagonal matrices $D$ with $D_{ii}\in \text{I}$. We propose a two-level nested methodology: an inner level where, for a given $\text{I}$, we compute an optimizer $D^\star(\text{I})$ by a gradient system approach, and an outer level where we tune $\text{I}$ so that the value $c$ is reached by $\mu(D^\star(\text{I})A)$. We extend the proposed two-level approach to the general multilayer, and possibly time-dependent, case $\dot{u}(t) = \sigma( A_k(t) \ldots \sigma ( A_{1}(t) u(t) + b_{1}(t) ) \ldots + b_{k}(t) )$ and we propose several numerical examples to illustrate its behaviour, including its stabilizing performance on a one-layer neural ODE applied to the classification of the MNIST handwritten digits dataset.