FieldSeer I: Physics-Guided World Models for Long-Horizon Electromagnetic Dynamics under Partial Observability

TL;DR

FieldSeer I is a physics-guided world model that accurately predicts electromagnetic field dynamics from partial observations, enabling interactive digital twins for photonic design with geometry modifications without re-assimilation.

Contribution

The paper introduces a novel geometry-aware world model that forecasts electromagnetic fields from partial data, incorporating physics guidance and enabling geometry edits without re-training.

Findings

Outperforms GRU and deterministic baselines in suffix fidelity

Achieves high accuracy in software-in-the-loop filtering and offline rollouts

Enables geometry modifications without re-assimilation

Abstract

We introduce FieldSeer I, a geometry-aware world model that forecasts electromagnetic field dynamics from partial observations in 2-D TE waveguides. The model assimilates a short prefix of observed fields, conditions on a scalar source action and structure/material map, and generates closed-loop rollouts in the physical domain. Training in a symmetric-log domain ensures numerical stability. Evaluated on a reproducible FDTD benchmark (200 unique simulations, structure-wise split), FieldSeer I achieves higher suffix fidelity than GRU and deterministic baselines across three practical settings: (i) software-in-the-loop filtering (64x64, P=80->Q=80), (ii) offline single-file rollouts (80x140, P=240->Q=40), and (iii) offline multi-structure rollouts (80x140, P=180->Q=100). Crucially, it enables edit-after-prefix geometry modifications without re-assimilation. Results demonstrate that geometry-conditioned world models provide a practical path toward interactive digital twins for photonic design.