{"id":"arxiv/2607.23143","url":"https://arxiv.org/abs/2607.23143","tags":["math.NA","physics.flu-dyn"],"score":0,"title":"Physics-informed token transformer methodology for nonlinear balance laws. I. Schwarzschild--Burgers fluid flows","domain":"math.NA","source":"arxiv","authors":["Philippe G. LeFloch","Shuyang Xiang"],"reviews":0,"subject":"math.NA","summary":"We introduce a Physics-Informed Token Transformer (PITT) methodology for nonlinear hyperbolic balance laws in one space dimension, using piecewise steady-state profiles for the representation of approximate weak solutions. The method combines symbolic equation tokenization, a Fourier neural operator encoder, an explicit Rankine--Hugoniot law for shock motion, and a learned correction term. For clarity, we present it here for the relativistic Schwarzschild--Burgers equation, a scalar model for spherically symmetric fluid flows on a Schwarzschild background. For this model the steady-state…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-25","reactions":{}}