Inverse design of metasurfaces using deep learning neural networks

使用深度学习神经网络进行超表面逆向设计

• 批准号: 2888733
• 金额: --
• 依托单位: University of Southampton
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2888733
• 关键词: Inverse; design; metasurfaces; using; deep

Deep learning artificial neural networks have seen a remarkable development in recent years and hold promise for a variety of applications in the field of nanophotonics. In this project the student will explore the use of deep learning techniques in nanophotonic systems at the frontiers of the field. Techniques of interest include the use of Physics informed networks to improve the quality of solutions, as well as recently proposed neural-adjoint optimization schemes. The student will explore how these techniques can be applied to current topics in optical metasurfaces and the study of light-matter coupling in hybrid quantum-electromagnetic systems. Theoretical and computational work will be complemented with proof of principle experiments involving the design of nanosystems that can be fabricated in our cleanroom facilities and their subsequent optical studies.

近年来，深度学习人工神经网络已经取得了显着的发展，并对纳米素养领域的各种应用保持了希望。在这个项目中，学生将探索该领域边界中纳米光系统中深度学习技术的使用。感兴趣的技术包括使用物理知情网络来提高解决方案的质量，以及最近提出的神经选择优化方案。学生将探讨如何将这些技术应用于光学跨面中的当前主题，并研究混合量子 - 电子磁系统中的光耦合。理论和计算工作将与主要实验的证据相辅相成，这些实验涉及纳米系统的设计，这些纳米系统可以在我们的洁净室设施中及其随后的光学研究进行捏造。