Deep learning for accelerated all-dielectric metasurface design.

By Christian C. Nadell, Bohao Huang, Jordan M. Malof, Willie J Padilla. 2019

316 citations recorded by the source corpus.

Deep learning has risen to the forefront of many fields in recent years, overcoming challenges previously considered intractable with conventional means. Materials discovery and optimization is one such field, but significant challenges remain, including the requirement of large labeled datasets and one-to-many mapping that arises in solving the inverse problem. Here we demonstrate modeling of complex all-dielectric metasurface systems with deep neural networks, using both the metasurface geometry and knowledge of the underlying physics as inputs. Our deep learning network is highly accurate, achieving an average mean square error of only 1.16 × 10-3 and is over five orders of magnitude faster than conventional electromagnetic simulation software. We further develop a novel method to solve the inverse modeling problem, termed fast forward dictionary search (FFDS), which offers tremendous controls to the designer and only requires an accurate forward neural network model. These techniques significantly increase the viability of more complex all-dielectric metasurface designs and provide opportunities for the future of tailored light matter interactions.

Subfields: Aerospace Engineering, Electrical and Electronic Engineering, Electronic, Optical and Magnetic Materials

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Chiara Daraio
2. Yongmin Liu
3. Peter R. Wiecha
4. Yun Raymond Fu
5. Cynthia Rudin
6. O. Muskens

In Electrical and Electronic Engineering

1. Goëry Genty
2. John M. Dudley
3. Sergei K. Turitsyn
4. Daniel Brunner
5. Ravi S. Prasher
6. Peter R. Wiecha
7. Sean D. Lubner
8. Yun Raymond Fu

In Electronic, Optical and Magnetic Materials

1. Yihao Xu
2. Chiara Daraio
3. Yongmin Liu
4. Ravi S. Prasher
5. Peter R. Wiecha
6. O. Muskens

Citations in the corpus

• How to See Hidden Patterns in Metamaterials with Interpretable Machine Learning (2021, methodology): “[6] Christian C Nadell, Bohao Huang, Jordan M Malof, Willie J Padilla, Deep learning for accelerated all-dielectric metasurface design, Opt.”
• Deep learning meets nanophotonics: A generalized accurate predictor for near fields and far fields of arbitrary 3D nanostructures. (2019, background): “68 Despite these limitations, our results clearly show that the ANN’s capabilities go far beyond that of a simple fit function, as it is able to apply trained behavior to new highly nontrivial con-figurations (see also Supporting Information Figure S8 on the prediction of curved structures).”
• Interfacing photonics with artificial intelligence: an innovative design strategy for photonic structures and devices based on artificial neural networks (2021, methodology): “In addition to spectrum prediction [55,56], the FCLs-based ANNs have also been used in the inverse design to realize other functionalities and benefit real-world applications [57 – 62].”
• Interpretable inverse design of particle spectral emissivity using machine learning (2020, methodology): “…machine learning has also shown some success in further accelerating the inverse design process by performing the inverse mapping in a single-shot (i.e. without iteratively solving the forward problem) for some optics problems (e.g., concentric spherical shells 12 and periodic metasurfaces 13–16 ).”
• Machine learning and applications in ultrafast photonics (2020, background): “Particular areas of success include in the design of nanomaterials and structures with specific target properties 9–11 , label-free cell classification 12 , super-resolution microscopy 13,14 , quantum optics 15 and optical communications 16–18 .”