STABILITY ANALYSIS OF NEURAL DIFFERENTIAL NETWORKS

STABILITY ANALYSIS OF NEURAL DIFFERENTIAL NETWORKS

Written by:ŞEVVAL HATİCE ÖTER

Department of Mathematics UNDERGRADUATE THESIS

Thesis Advisor:Prof. Canan ÇELİK KARAASLANLI

İSTANBUL, 2025

This thesis was primarily written in LaTeX. Since it is difficult to display mathematical formulas and other structures on this blog, screenshots from the PDF have been included.


where:

• Its time derivative is non-positive.

Thus, the system satisfies the criteria for Lyapunov stability. This stability analysis demon-

strates that the neural differential equation model converges to a stable equilibrium, making it

suitable for applications requiring robust and predictable dynamics

Share: