Id: arxiv/2611.14650 Url: https://arxiv.org/abs/2611.14650 Tags: gr-qc, hep-th Score: 0 Title: Ernst-Potential Driven Recovery of Kerr with Neural Networks Domain: gr-qc Source: arxiv Authors: Veselin G. Filev, Stoytcho S. Yazadjiev Reviews: 0 Subject: gr-qc Category: research Comments: 0 Disputed: false Division: physics Mentions: 0 Discussed: 0 Published: 2026-11-10 Page: https://delphi.science/r/arxiv/2611.14650 We use artificial neural networks to recover the Kerr solution by directly minimizing a regularized Ernst sigma-model action. Reduction along the axial Killing direction avoids the divergence at the ergosphere. Compactified coordinates cover the entire black-hole exterior, while the network parametrization enforces the asymptotic data and axis regularity. For dimensionless spins $0.1 \le a/M \le 0.99$, the median maximum relative error in the recovered metric component $g_{φφ}$ remains below $7.2\times10^{-4}$. When only the horizon rod length and angular momentum are prescribed, the…