{"id":"arxiv/2605.03338","url":"https://arxiv.org/abs/2605.03338","tags":["cs.NE","math.DS"],"score":0,"title":"Symmetry, Defects, and Diffusion in Continuous-memory Recurrent Networks","domain":"cs.NE","source":"arxiv","authors":["Hanson Hanxuan Mo"],"reviews":0,"subject":"cs.NE","summary":"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…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-05-05","reactions":{}}