Many physical systems exhibit one-way coupling: exogenous fields such as medium density or obstacle geometry influence the dynamics without being affected by them.
We investigate how explicitly representing these fields can improve the accuracy and computational efficiency of transformer-based PDE surrogates.
Introduction
Building on spatially adaptive normalization, we introduce Patchwise Adaptive Normalization (PAdaNorm), which partitions exogenous fields into patches aligned with the state tokenization.
We evaluate PAdaNorm on datasets from three different domains, using a Walrus-inspired transformer architecture.
Results
Across the three PDE families, PAdaNorm is on par with or up to 16% better than parameter-matched baselines in mean long-horizon variance-scaled RMSE on 60-step autoregressive rollouts.
When the conditioning is static, PAdaNorm caches the conditioning embeddings once and reuses them across the entire rollout, reducing inference FLOPs per step by up to 36% compared with parameter-matched baselines.
An ablation over eight conditioning variants shows that modulating a single attention axis (temporal or spatial) suffices and that a lightweight shared MLP across blocks remains competitive with per-block projections.