Multifunctional Metasurface Design via Physics‐Simplified Machine Learning

International Journal of Intelligent Systems, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Metasurface can manipulate electromagnetic (EM) waves flexibly, which provides the basis for functional integration. Recently, efficient machine‐learning‐assisted methods have attracted intensive attentions in multifunctional metasurfaces design. However, conventional design is to fit internal relationship form of black box, ignores underlying physical logic, resulting increased complexity machine learning architecture with parameters increasing. In order adapt multiparameter optimization design, we propose a multiplexing neural network (MNN) based on decoupling at layer simplify both structural and architecture. The four interacting are simplified into independently regulated so that facile functions be realized only by simple network. For verification, scattering, anomalous reflection, focusing, hologram integrated same metasurface aperture MNN. Performances fully demonstrated simulation measurement. Importantly, this work paves way bidirectional simplification via inspiration, an method potentially applied satellite communications other fields.

Язык: Английский

---

Advances in artificial intelligence for artificial metamaterials

APL Materials, Год журнала: 2024, Номер 12(12)

Опубликована: Дек. 1, 2024

The 2024 Nobel Prizes in Physics and Chemistry were awarded for foundational discoveries inventions enabling machine learning through artificial neural networks. Artificial intelligence (AI) metamaterials are two cutting-edge technologies that have shown significant advancements applications various fields. AI, with its roots tracing back to Alan Turing’s seminal work, has undergone remarkable evolution over decades, key including the Turing Test, expert systems, deep learning, emergence of multimodal AI models. Electromagnetic wave control, critical scientific research industrial applications, been significantly broadened by metamaterials. This review explores synergistic integration metamaterials, emphasizing how accelerates design functionality materials, while novel physical networks constructed from enhance AI’s computational speed ability solve complex problems. paper provides a detailed discussion AI-based forward prediction inverse principles metamaterial design. It also examines potential big-data-driven methods addressing challenges In addition, this delves into role advancing focusing on progress electromagnetic optics, terahertz, microwaves. Emphasizing transformative impact intersection between underscores improvements efficiency, accuracy, applicability. collaborative development process opens new possibilities innovations photonics, communications, radars, sensing.

Язык: Английский

---

Artificial Intelligence and Machine Learning for materials

Current Opinion in Solid State and Materials Science, Год журнала: 2024, Номер 34, С. 101202 - 101202

Опубликована: Окт. 9, 2024

Язык: Английский

---