{"id":"arxiv/2610.24202","url":"https://arxiv.org/abs/2610.24202","tags":["math.NA","cs.LG","math.FA"],"score":0,"title":"Certified neural approximations for parabolic obstacle problems","domain":"math.NA","source":"arxiv","authors":["Davood Damircheli"],"reviews":0,"subject":"math.NA","summary":"Residual-trained neural surrogates carry no error guarantee, a limitation for free-boundary problems such as phase change and contact. We derive guaranteed error bounds for neural approximations of parabolic obstacle problems in first-order state--flux--multiplier form. The main result bounds the flux error of the network and the energy error of its feasible projection $\\max\\{v,g\\}$ by the constitutive residual, the initial mismatch, the complementarity defect, and the PDE residual weighted by $c_F^2/α$, without the coercivity and norm-equivalence constants of the first-order formulation. It…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-16","reactions":{}}