Out-of-distribution transfer of PDE foundation models to material dynamics under extreme loading

TL;DR

This paper evaluates the ability of PDE foundation models, pretrained on fluid data, to transfer to extreme material dynamics involving shocks and fractures, highlighting their performance and sample efficiency in out-of-distribution scenarios.

Contribution

It introduces a benchmark for out-of-distribution transfer of PDE models to material dynamics and compares pretrained models with training from scratch under distribution shifts.

Findings

Pretrained models show varying transfer capabilities to shock-driven regimes.

Fine-tuning pretrained models improves sample efficiency over training from scratch.

The benchmark provides a standardized evaluation protocol for PDE model transferability.

Abstract

Most PDE foundation models are pretrained and fine-tuned on fluid-centric benchmarks. Their utility under extreme-loading material dynamics remains unclear. We benchmark out-of-distribution transfer on two discontinuity-dominated regimes in which shocks, evolving interfaces, and fracture produce highly non-smooth fields: shock-driven multi-material interface dynamics (perturbed layered interface or PLI) and dynamic fracture/failure evolution (FRAC). We formulate the downstream task as terminal-state prediction, i.e., learning a long-horizon map that predicts the final state directly from the first snapshot without intermediate supervision. Using a unified training and evaluation protocol, we evaluate two open-source pretrained PDE foundation models, POSEIDON and MORPH, and compare fine-tuning from pretrained weights against training from scratch across training-set sizes to quantify sample efficiency under distribution shift.