{"id":"arxiv/2604.05967","url":"https://arxiv.org/abs/2604.05967","tags":["cs.LG","math.DS","math.OC"],"score":0,"title":"On Dominant Manifolds in Reservoir Computing Networks","domain":"cs.LG","source":"arxiv","authors":["Noa Kaplan","Alberto Padoan","Anastasia Bizyaeva"],"reviews":0,"subject":"cs.LG","summary":"Understanding how training shapes the geometry of recurrent network dynamics is a central problem in time-series modeling. We study the emergence of low-dimensional dominant manifolds in the training of Reservoir Computing (RC) networks for temporal forecasting tasks. For a general linear continuous-time reservoir in the infinite-data limit, we show that the training data generate an invariant subspace of the trained reservoir, whose dimension equals the number of dominant modes. We then specialize to a simplified diagonal linear reservoir, where we link the dominant eigenvalues and…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-04-07","reactions":{}}