Transferability of Learned States in Neural PDE Solvers

A literature audit of 12 papers produced 18 version-specific protocol records showing that reuse in neural PDE solvers cannot be judged by final accuracy alone, according to an arXiv paper (2610.10972v1). Across 240 source-training trajectories spanning two linear PDE families, Fourier neural operators and convolutional networks, a fixed predictor's benefit reversed across correction algorithms, and work-based selection saved 2.50-3.33 conjugate-gradient iterations on held-out in-distribution tasks. Independent batches confirmed a 0.73 percent complete online saving for one physics-trained Fourier neural operator against zero-initialized Poisson-preconditioned CG.

arXiv:2610.10972v1 Announce Type: new Abstract: Assessing useful reuse in neural PDE solvers is challenging: final accuracy can reflect source learning and target-time computation. Our reuse contract separates solution accuracy, learning contribution, and numerical utility through paired state comparisons, matched target information and budgets, and cost accounting. A literature audit extracts 18 version-specific protocol records from 12 papers, documenting retained states, target-time resources, and reported controls. For a fixed linear system and residual tolerance, we construct two initial guesses with identical solution-error, energy-error, and residual norms, reaching the same solution with different conjugate-gradient CG iteration counts. Across 240 source-training trajectories, two linear PDE families, Fourier neural operators and convolutional networks, a fixed predictor's benefit reverses across correction algorithms. Among pairs with both relative prediction errors less than or equal to 5 percent on 64 in-distribution tasks 63 by 63 interior grids , reductions in all three norms accompany more CG iterations, at mean taskwise rates of 23.5 percent and 23.9 percent in two libraries. Work-based selection saves 2.50-3.33 CG iterations on held-out in-distribution tasks; matched adaptation demonstrates finite-budget pretraining value. Independent batches confirm a 0.73 percent complete online saving for one physics-trained Fourier neural operator against zero-initialized Poisson-preconditioned CG. Reuse requires matched state comparisons and downstream computational evidence.