AI 中文总结
该研究提出将非局域性作为拓扑光子学的设计参数，通过校准工作流程明确拓扑设计准则的可靠性，利用非局域耦合实现鲁棒纳米光子器件。
AI 中文摘要
拓扑光子学已提供了强大的设计准则，可通过几何、对称性和工程化耦合实现光的路由与局域化。然而，制备的纳米光子学和等离激元器件极少仅实现紧凑拓扑模型所假设的局域或短程相互作用网络。长程近场耦合、 retardation（延迟效应）、辐射泄漏、衬底辅助杂化、材料色散和制备无序可重塑被视为拓扑的光学模式。在本观点中，我主张应将非局域性视为设计参数而非残余微扰。这一转变要求从理想相位标签转向校准的相互作用模型、有限结构可观测量、鲁棒性图谱和分级置信度度量。我讨论了全波模拟、实验及物理引导的机器学习如何将几何设计变量与有效电磁相互作用网络关联。此类校准工作流程可阐明拓扑设计准则何时可靠、何时失效，以及如何利用非局域耦合实现鲁棒纳米光子器件。
英文摘要
Topological photonics has provided powerful design rules for routing and localizing light through geometry, symmetry and engineered coupling. However, fabricated nanophotonic and plasmonic devices rarely realize only the local or short-range interaction networks assumed in compact topological models. Long-range near-field coupling, retardation, radiation leakage, substrate-assisted hybridization, material dispersion and fabrication disorder can reshape the optical modes that are interpreted as topological. In this Perspective, I argue that nonlocality should be treated as a design parameter rather than as a residual perturbation. This shift requires moving from ideal phase labels toward calibrated interaction models, finite-structure observables, robustness maps and graded confidence measures. I discuss how full-wave simulations, experiments and physics-informed learning can connect geometric design variables to effective electromagnetic interaction networks. Such calibrated workflows can clarify when a topological design rule is reliable, when it fails, and how nonlocal coupling can be exploited for robust nanophotonic devices.
