Id: arxiv/2607.00237 Url: https://arxiv.org/abs/2607.00237 Tags: math.AP, math.AC Score: 0 Title: Tropical Geometry as a Restricted Architecture for Physics-Informed Neural Networks: Applications in Nonlinear Fluid-Structure Examples Domain: math.AP Source: arxiv Authors: Carla Valencia-Negrete, Cristhian Garay-Lopez, Marco Favela-Rodriguez, Alonso Andapia-Viveros Reviews: 0 Subject: math.AP Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-06-30 Page: https://delphi.science/r/arxiv/2607.00237 Nonlinear algebraic (polynomial) differential equations that govern fluid-structure interactions, such as those modeling vortex-induced vibrations, and shock waves, often lack analytical solutions, creating significant challenges to efficient prediction and control. While Physics-Informed Neural Networks (PINNs) offer a mesh-free numerical alternative, they frequently suffer from convergence stagnation when optimizing over chaotic landscapes or stiff singularities. This paper introduces a hybrid methodology that integrates tropical differential algebraic geometry with deep learning. Using…