Id: arxiv/2611.18992 Url: https://arxiv.org/abs/2611.18992 Tags: math.NA, math.OC Score: 0 Title: NeRI: Neural Residual Iteration for Elliptic Operator Equations and Variational Inequalities Domain: math.NA Source: arxiv Authors: Michael Hintermüller, Jianfeng Ning Reviews: 0 Subject: math.NA Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-11-13 Page: https://delphi.science/r/arxiv/2611.18992 In this work, we propose a so-called {\it Neural Residual Iteration} (NeRI), a weak-form neural-network framework for solving a class of elliptic operator equations and variational inequalities, including nonlinear and nonsymmetric problems that may not admit a global energy functional. The key idea is to construct, at each iteration, a state-dependent linear functional whose derivative with respect to the trial function recovers the current weak residual. A single gradient step is then performed on the residual, yielding a residual-driven neural update without requiring a global energy…