SOURCE-LINKED INTELLIGENCE

How Does Distribution Shift Shape Pretraining Gains in Neural PDE Surrogates?

arXiv · AI, language, vision and robotics · article · Sep 17, 2026 · UTC

Pretraining a neural PDE surrogate can reduce the amount of new CFD data needed when geometry or modeled physics changes. However, it remains unclear how different components of distribution shift affect this benefit. We pretrain a surrogate on 254,909 RANS solutions from one airfoil family and fine-tune it on a new family under two target settings with matched freestream ranges: the same Spalart-Allmaras (SA) modeling and SA with added $e^N$ transition modeling. At $N=1000$, the pretrained model matches the accuracy of a model trained from scratch on $3.25\times$ as many samples for the same-

Read original source ↗ Open in workspace

recordType

paper

region

Global