Skip to content
Preprint

Low Residuals, Wrong Value: Learning the Viscosity Solution of Hamilton-Jacobi-Isaacs Equations with PINNs

Oct 2026 · 0 citations · 55 references
Mathematics Computer Science

TL;DR

This work uses the semiconcavity of the viscosity solution to build a mesh-free finite-difference test that needs only forward network evaluations and detects the folded solutions and the harmful min-of-charts failures without access to the exact value function.

Abstract

Physics-informed neural networks (PINNs) and other residual-trained solvers judge a candidate solution by the size of its sampled PDE residual. For first-order Hamilton-Jacobi equations this is not sufficient: the same equation and boundary data admit Lipschitz almost-everywhere solutions that differ from the viscosity solution by $O(1)$ while violating the equation only on sets of measure zero. We study whether residual minimization selects such solutions in a fixed-horizon pursuit-evasion game whose Hamilton-Jacobi-Isaacs value function is known in closed form on the circle $S^1$ and on the flat tori $\mathbb{T}^{2}$ and $\mathbb{T}^{4}$. We train four neural parameterizations with ten seeds on each geometry, 120 runs in total. A standard network with periodic Fourier features converges to folded almost-everywhere solutions in all 30 runs: small residuals and the expected capture-layer gradient, yet values that are wrong over most of the domain. A network with periodicity imposed through the loss recovers the viscosity solution in one and two dimensions but becomes unreliable in four. Two parameterizations that receive the geodesic distance do substantially better: a single network fed the wrapped displacement and its distance recovers the viscosity solution in all 30 runs, while a minimum over periodic image charts succeeds in 27 of 30 and introduces a wrong-image failure mode. Finally, for the convex Hamiltonian considered here, we use the semiconcavity of the viscosity solution to build a mesh-free finite-difference test that needs only forward network evaluations; it detects the folded solutions and the harmful min-of-charts failures without access to the exact value function. A small sampled residual thus does not by itself validate a learned Hamilton-Jacobi solution; the parameterization and independent structural diagnostics can be essential for selecting the viscosity solution.

View source

Similar papers

#machine learning Preprint Oct 2026

Convex Equilibrium Potentials for Monge--Amp\`ere: Curvature Bounds and A Posteriori Evaluation

Equilibrium neural networks define their hidden state as the fixed point of a contraction, so their effective depth is set at evaluation time rather than by the architecture. Using them for fully nonlinear PDEs raises two issues that explicit networks avoid: second derivatives may grow with the effective depth, and the...

E. Vázquez, Ying-Ling Lu, T. Nowicki · 0 citations
Open access Nov 2026

Neural Finite Volume and Mixture-of-Experts Physics-Informed Networks for the Isentropic Cargo–LeRoux Model: An Exact-Solution-Grounded Comparative Study

We compare three neural approaches to the Riemann problem of the isentropic Cargo–LeRoux model with an exact solver and with classical Godunov-type schemes. The model has a linearly degenerate contact field and a strictly convex closure that is singular at aρ=1. We pose the star-state system as constrained residual min...

Saleem Ahmed Naich, Sidrah Ahmed · 0 citations
Preprint Sep 2026

Prediction with Five Experts and Geometric Stopping: A Probabilistic Construction and Analytic Verification

We present a solution of the limiting five-expert prediction problem with geometric stopping, based on stochastic calculus and partial differential equations. If $\delta$ is the stopping probability, the minimax expected regret from tied initial scores is $45\pi^2/(512\sqrt{2\delta})+o(\delta^{-1/2})$ as $\delta\downar...

Erhan Bayraktar, Ibrahim Ekren, N. Kolliopoulos · 3 citations
Preprint Oct 2026

Loss of Monotonicity in the Vector Play Model and Consequences for Newton-Type Solvers

The vector play model (VPM) replaces the implicit minimization of the energy-based hysteresis model (EBM) by an explicit update. Its tangent contains the product of the Hessians of two unrelated potentials and is unsymmetric. Hence, the VPM has no co-energy density at frozen memory, and the skew part makes $\boldsymbol...

Dominik Cibicki, K. Roppert, M. Kaltenbacher · 0 citations
Preprint Oct 2026

S$^{2}$-PINN: Stochastic Separable Physics-Informed Neural Networks

Uncertainty quantification (UQ) for random partial differential equations (PDEs) is ubiquitous in computational science and engineering. However, classical spectral solvers for this class of problems face the curse of dimensionality, and existing neural solvers often ignore the stochastic structure that makes moments a...

Zhendong Li, Akwum Onwunta · 0 citations
Preprint Oct 2026

A Method for Linearization of Systems of Quadratic Equations

Integrating physical conservation laws into deep neural architectures remains a central SciML challenge. Soft-penalty approaches, including many PINNs, may be sensitive to loss weighting and optimization; generic DEQ and implicit-layer solvers may require instance-dependent iteration counts. We study a norm-linearizati...

Анатолий Пак, A. Pak · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.