Measuring and Mitigating Forgetting when Extending a PDE Foundation Model to Simulated Flow Regimes
Year: 2026
Type: preprint
Abstract: Foundation models for partial differential equations adapt efficiently to new tasks, but their effect on the original pretraining distribution is rarely measured. We study this retention problem on Poseidon, a neural operator pretrained on periodicdomain compressible Euler and incompressible Navier–Stokes families. We adapt it to two regimes absent from pretraining, a compressible Kelvin–Helmholtz flow and wall-bounded flow past a circular cylinder. Fine-tuning adapts efficiently but degrades the pretraining families. LoRA still forgets substantially, and freezing the whole trunk reduces forgetting without preventing it. We show that low-weight replay and joint training resolve this trade-off, extending the model to the new regimes while keeping the pretraining families near their original accuracy. These results suggest that lightweight replay can extend a PDE foundation model while substantially preserving its existing capabilities.
Source: Research Square
Authors: wasif masood, Shoaib Noor
Language: English
Subfield: Statistical and Nonlinear Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
Award: G9001396210
OpenAlex ID: W7222918237
