{"id":"arxiv/2606.21828","url":"https://arxiv.org/abs/2606.21828","tags":["math.NA","cs.LG"],"score":0,"title":"Spectrally Safe Neural Operator Warm-Starts for Large-Scale Newton Solvers","domain":"math.NA","source":"arxiv","authors":["Jaemin Oh","Youngkyu Lee","Jerome Darbon","George Em Karniadakis"],"reviews":0,"subject":"math.NA","summary":"Neural operators are increasingly used to warm-start Newton solvers for nonlinear PDEs, on the premise that a low test error places the initial guess inside the basin of attraction. We show that this premise is unreliable. An operator trained to the relative \\(L^2\\) error \\(O(10^{-3})\\) can still produce an initial state in which the discrete Jacobian is indefinite, because the mean-squared training controls error on average while leaving localized pointwise violations of the underlying physics. For a nearly incompressible hyperelasticity problem, we trace this to the predicted volume…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-20","reactions":{}}