Deep Neural Operator Enabled Concurrent Multitask Design for Multifunctional Metamaterials Under Heterogeneous Fields

• Published in: Advanced optical materials Vol. 12; no. 15
• Main Authors: Lee, Doksoo, Zhang, Lu, Yu, Yue, Chen, Wei
• Format: Journal Article
• Language: English
• Published: Weinheim Wiley Subscription Services, Inc 01.05.2024
• Subjects: concurrent optimization; deep learning for photonics; heterogeneous fields; Inverse design; Inverse problems; Metamaterials; multifunctional metamaterials; Nanoantennas; neural operator; Phase distribution
• ISSN: 2195-1071; 2195-1071
• DOI: 10.1002/adom.202303087

Multifunctional metamaterials (MMM) bear promise as next‐generation material platforms supporting miniaturization and customization. Despite many proof‐of‐concept demonstrations and the proliferation of deep learning assisted design, grand challenges of inverse design for MMM, especially those involving heterogeneous fields possibly subject to either mutual meta‐atom coupling or long‐range interactions, remain largely under‐explored. To this end, a data‐driven design framework is presented, which streamlines the inverse design of MMMs involving heterogeneous fields. A core enabler is implicit Fourier neural operator (IFNO), which predicts heterogeneous fields distributed across a metamaterial array, thus in general at odds with homogenization assumptions. Additionally, a standard formulation of inverse problem covering a broad class of MMMs is presented, together with gradient‐based multitask concurrent optimization identifying a set of Pareto‐optimal architecture‐stimulus (A‐S) pairs. Fourier multiclass blending is proposed to synthesize inter‐class meta‐atoms anchored on a set of geometric motifs, while enjoying training‐free dimension reduction and built‐it reconstruction. Interlocking the three pillars, the framework is validated for light‐by‐light programmable nanoantenna, whose design involves vast space jointly spanned by quasi‐freeform supercells, maneuverable incident phase distributions, and conflicting figure‐of‐merits (FoM) involving on‐demand localization patterns. Accommodating all the challenges, the framework can propel future advancements of MMM. A data‐driven design framework powered by neural operator for multifunctional metamaterials is proposed. It interlocks a field‐to‐field surrogate using Fourier neural operators, concurrent multitask optimization for Pareto‐optimality, and Fourier multiclass blending for meta‐atom synthesis. The framework simultaneously handles the joint architecture‐stimulus space, long‐range interactions, and trade‐offs among functionalities. Its efficacy is demonstrated for concurrent design of spatially addressable metasurfaces.