Full Lyapunov Spectra of Random Neural Networks, Part I: Finite-Time and Asymptotic Theory
Abstract
We develop a dynamical mean-field framework for the full finite-time and stationary asymptotic Lyapunov spectra of random recurrent neural networks. At finite observation time, an exact log-determinant representation of tangent-propagator singular values leads to a closed-contour matrix Dyson equation for the smoothed spectral tail. In the stationary setting, a cocycle representation yields an operator equation for the asymptotic cumulative spectrum. Both formulations retain complete local gain histories and the repeated action of quenched connectivity. Direct singular-value calculations and network QR spectra support the predictions in the regimes examined, with sampling uncertainty and numerical cutoffs assessed separately. The stationary equation also recovers the known maximal exponent, and controlled gain-history replay illustrates how temporal correlations redistribute instability. This work was carried out in collaboration with ChatGPT 5.6 Sol. Toward the end of preparation we became aware of David G. Clark's contemporaneous work, \emph{Lyapunov spectrum of random neural networks}[arXiv:2610.12426], which reports overlapping mean-field results for the full Lyapunov spectrum using a cavity approach.
Related
- Lyapunov spectrum of random neural networks
- Dynamics of random recurrent neural networks: colored noise and sublinear activations
- Nonlinear dynamics of random neural networks with second-order synaptic motifs
- Stationary covariance spectra of discrete-time non-normal random recurrent dynamics
- SPECTRA: predicting cellular perturbation responses with Graph Learning over Gene Regulatory Networks
BibTeX
@misc{wang2026full,
title = {{Full Lyapunov Spectra of Random Neural Networks, Part I: Finite-Time and Asymptotic Theory}},
author = {Haoyu Albert Wang and Wei P. Dai},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.14922},
eprint = {2610.14922},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.14922},
}
RIS
TY - UNPB
TI - Full Lyapunov Spectra of Random Neural Networks, Part I: Finite-Time and Asymptotic Theory
AU - Haoyu Albert Wang
AU - Wei P. Dai
PY - 2026
DA - 2026/10/10
PB - arXiv
DO - 10.48550/arxiv.2610.14922
UR - https://arxiv.org/abs/2610.14922
AB - We develop a dynamical mean-field framework for the full finite-time and stationary asymptotic Lyapunov spectra of random recurrent neural networks. At finite observation time, an exact log-determinant representation of tangent-propagator singular values leads to a closed-contour matrix Dyson equation for the smoothed spectral tail. In the stationary setting, a cocycle representation yields an operator equation for the asymptotic cumulative spectrum. Both formulations retain complete local gain histories and the repeated action of quenched connectivity. Direct singular-value calculations and network QR spectra support the predictions in the regimes examined, with sampling uncertainty and numerical cutoffs assessed separately. The stationary equation also recovers the known maximal exponent, and controlled gain-history replay illustrates how temporal correlations redistribute instability. This work was carried out in collaboration with ChatGPT 5.6 Sol. Toward the end of preparation we became aware of David G. Clark's contemporaneous work, \emph{Lyapunov spectrum of random neural networks}[arXiv:2610.12426], which reports overlapping mean-field results for the full Lyapunov spectrum using a cavity approach.
ER -