Symmetry, Defects, and Diffusion in Continuous-memory Recurrent Networks
Abstract
Continuous-memory recurrent networks must preserve phase, position, or orientation despite model imperfections and state noise. We develop a geometric framework that separates three questions: how many memory coordinates are neutrally transported, how deterministic perturbations alter their finite-horizon stability, and how ambient noise is decoded along them. Exact per-input equivariance transports analytical group tangents pathwise and, on a compact nondegenerate orbit stratum, yields at least q=dim(G/H) zero group-tangent Lyapunov exponents under stationary ergodic driving. For imperfect dynamics, a four-block tangent/normal decomposition gives local and finite-horizon bounds on tangent growth and subspace rotation, distinguishing first-order direct damage from second-order leakage through contracting normal directions. For noisy dynamics, a specified decoder maps ambient covariance Q to coordinate covariance ZQZ⊤; under isotropic noise and fixed tangent energy, least-squares decoding and scaled-isometric action geometry minimize local diffusion. Local and finite-horizon evaluations include cases both within and outside the sufficient conditions. In a fresh twenty-seed T2 replication, a decoder-covariance objective improves noisy horizon-256 memory in every pair while meeting a prespecified clean-error equivalence margin. Direct noise training also improves noisy memory but incurs a clean-error tradeoff. In coupled T4/T8 integrators, the advantage of anisotropic-covariance over isotropic regularization reverses when evaluation noise becomes isotropic. These results connect continuous symmetry to measurable limits and design choices for recurrent memory under explicit dynamical, decoder, and noise assumptions.
Related
- Some new numerical schemes for delay sub-diffusion equations via the idea of symmetry
- Validation of Symmetry Induced Bifurcation Points in Nonlinear Diffusion Problems
- Continuous Diffusion Scales Competitively with Discrete Diffusion for Language
- Continuous-time q-Learning for Finite-Horizon Portfolio Choice with Optimal Stopping in Jump-Diffusion Markets
- Disentangling Continuous-Time Latent Dynamics: Identifiability of Latent SDEs via Diffusion Shifts
BibTeX
@misc{mo2026symmetry,
title = {{Symmetry, Defects, and Diffusion in Continuous-memory Recurrent Networks}},
author = {Hanson Hanxuan Mo},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2605.03338},
eprint = {2605.03338},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.03338},
}
RIS
TY - UNPB
TI - Symmetry, Defects, and Diffusion in Continuous-memory Recurrent Networks
AU - Hanson Hanxuan Mo
PY - 2026
DA - 2026/09/09
PB - arXiv
DO - 10.48550/arxiv.2605.03338
UR - https://arxiv.org/abs/2605.03338
AB - Continuous-memory recurrent networks must preserve phase, position, or orientation despite model imperfections and state noise. We develop a geometric framework that separates three questions: how many memory coordinates are neutrally transported, how deterministic perturbations alter their finite-horizon stability, and how ambient noise is decoded along them. Exact per-input equivariance transports analytical group tangents pathwise and, on a compact nondegenerate orbit stratum, yields at least q=dim(G/H) zero group-tangent Lyapunov exponents under stationary ergodic driving. For imperfect dynamics, a four-block tangent/normal decomposition gives local and finite-horizon bounds on tangent growth and subspace rotation, distinguishing first-order direct damage from second-order leakage through contracting normal directions. For noisy dynamics, a specified decoder maps ambient covariance Q to coordinate covariance ZQZ^⊤; under isotropic noise and fixed tangent energy, least-squares decoding and scaled-isometric action geometry minimize local diffusion. Local and finite-horizon evaluations include cases both within and outside the sufficient conditions. In a fresh twenty-seed T^2 replication, a decoder-covariance objective improves noisy horizon-256 memory in every pair while meeting a prespecified clean-error equivalence margin. Direct noise training also improves noisy memory but incurs a clean-error tradeoff. In coupled T^4/T^8 integrators, the advantage of anisotropic-covariance over isotropic regularization reverses when evaluation noise becomes isotropic. These results connect continuous symmetry to measurable limits and design choices for recurrent memory under explicit dynamical, decoder, and noise assumptions.
ER -