Inf-Sup Neural Networks for High Dimensional PDEs
arXiv:2607.11718
The paper introduces a neural‑network framework that rewrites high‑dimensional partial differential equations as inf‑sup (saddle‑point) optimization problems using a Lagrange multiplier, and provides theoretical equivalence, error estimates, and numerical results.
Abstract
Solving partial differential equations (PDEs) in high dimensions remains challenging due to the curse of dimensionality. We propose a neural-network-based framework that reformulates PDEs as inf--sup optimization problems through the introduction of a Lagrange multiplier. The primal solution and the associated Lagrange multiplier are parameterized by two networks and are computed via an iterative saddle-point optimization procedure. We prove the theoretical equivalence between the proposed optimization formulation and the original PDE problem, and we derive rigorous error estimates that quantify the total approximation error in terms of the network approximation error, statistical (sampling) error, and optimization error. Numerical experiments demonstrate the accuracy, stability, and efficiency of the proposed method for solving high-dimensional PDEs.