Nanophotonic particle simulation and inverse design using artificial neural networks

Peurifoy, John; Shen, Yichen; Li, Jing; Yang, Yi; Cano-Renteria, Fidel; DeLacy, Brendan G.; Joannopoulos, John D.; Tegmark, Max; Soljačić, Marin

doi:10.1126/sciadv.aar4206

Cited by 736 publications

(518 citation statements)

References 20 publications

Supporting

Mentioning

492

Contrasting

Order By: Relevance

“…An iterative process can be used to optimize the input structural parameters to accomplish the goal of the inverse design. [14] In contrast, in an inverse design network, [11] the desired optical properties are taken as the input, and the network directly outputs the device structures. This method eliminates the iterative optimization process.…”

Section: Doi: 101002/adma201905467mentioning

confidence: 99%

“…In the field of nanophotonics, computational inverse design can reshape the landscape and techniques available to complex and emerging applications . Recent advancements in deep neural networks (DNNs) have demonstrated efficient forward‐modeling that can predict resonance spectrum accurately, and perform the inverse design of photonic device structures . The general steps usually involve a one‐time investment of sufficient EM simulation data, which are composed of variable device parameters and corresponding optical resonance at different wavelengths, followed by constructing DNNs.…”

mentioning

confidence: 99%

“…In a forward modeling network, the input takes the structural parameters and the network outputs optical properties such as spectra. An iterative process can be used to optimize the input structural parameters to accomplish the goal of the inverse design . In contrast, in an inverse design network, the desired optical properties are taken as the input, and the network directly outputs the device structures.…”

mentioning

confidence: 99%

“…[9] Recent advancements in deep neural networks (DNNs) have demonstrated efficient forward-modeling that can predict resonance spectrum accurately, and perform the inverse design of photonic device structures. [10][11][12][13][14][15][16][17][18] The general steps usually involve a one-time investment of sufficient EM simulation data, which are composed of variable device parameters and corresponding optical resonance at different wavelengths, followed by constructing DNNs. In a forward modeling network, Silicon nanostructure color has achieved unprecedented high printing resolution and larger color gamut than sRGB.…”

mentioning

confidence: 99%

See 3 more Smart Citations

A Bidirectional Deep Neural Network for Accurate Silicon Color Design

et al. 2019

View full text Add to dashboard Cite

Silicon nanostructure color has achieved unprecedented high printing resolution and larger color gamut than sRGB. The exact color is determined by localized magnetic and electric dipole resonance of nanostructures, which are sensitive to their geometric changes. Usually, the design of specific colors and iterative optimization of geometric parameters are computationally costly, and obtaining millions of different structural colors is challenging. Here, a deep neural network is trained, which can accurately predict the color generated by random silicon nanostructures in the forward modeling process and solve the nonuniqueness problem in the inverse design process that can accurately output the device geometries for at least one million different colors. The key results suggest deep learning is a powerful tool to minimize the computation cost and maximize the design efficiency for nanophotonics, which can guide silicon color manufacturing with high accuracy.

show abstract

Section: Doi: 101002/adma201905467mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

A Bidirectional Deep Neural Network for Accurate Silicon Color Design

et al. 2019

View full text Add to dashboard Cite

show abstract

“…[15][16][17][18][19][20] To identify the optimized parameters of a device, the algorithm computes the gradient, or sensitivity, through the corresponding adjoint problem and updates the parameters along the deepestgradient direction. [27][28][29][30][31] In conjunction with traditional optimization techniques, it has been proved that deep learning can substantially mitigate problems such as the convergence to local minima and the curse of dimensionality in other optimization schema. [21][22][23] The philosophy of the algorithms is to treat photonic structures as a population of individuals, and carry out bio-inspired operations such as selection, reproduction, and mutation to the population in order to identify the optimized individual through evolution.…”

Section: Doi: 101002/adma201904790mentioning

confidence: 99%

Compounding Meta‐Atoms into Metamolecules with Hybrid Artificial Intelligence Techniques

et al. 2019

View full text Add to dashboard Cite

controllability of light due to the limited flexibility rendered in periodic metastructures of simple unit cells. To overcome these deficiencies, metasurfaces comprised of multiple meta-atoms, such as gradient and multilayered metasurfaces, have been proposed and developed. [7][8][9] Relying on the collective effects of multiple meta-atoms, these metasurfaces present intriguing properties such as anomalous deflection, [7,10] arbitrary phase control, asymmetric polarization conversion, [8,11] wave-front shaping, [12][13][14] etc., which brings about extensive applications for imaging, optical signal processing, emission control, and much more. Here in our following discussion, we refer to unit cells composed of various meta-atoms as metamolecules, analogous to the hierarchical relationship between atoms and molecules in nature. In our definition of a metamolecule, we assume every two adjacent meta-atoms are not strongly coupled, in which case the overall properties of the metamolecule can be analytically predicted by the properties of its constituent meta-atoms. Such an assumption is valid in most metasurfaces that consist discrete, spatially variant building blocks.Despite the extraordinary properties of metasurfaces made up of metamolecules, designing multiple meta-atoms that collectively function as a device is a time-consuming task that requires labor-intensive trial-and-error simulations. The difficulty of the inverse design of such metamolecules arises from the intricate mechanisms of multistructured systems, the vast number of possible combinations of distinct meta-atoms, as well as the expensive 3D full wave simulations required. Traditionally, a practical solution to such a design follows three steps: 1) specifying a class of geometry with a few parameters as candidate meta-atoms, 2) carrying out parametric sweeps on these parameters, and 3) enumerating possible combinations of meta-atoms to meet the design objective. However, the limitation of the geometry in the strategy largely restricts the variety of the shapes of meta-atoms, which usually does not lead to an optimal solution, even after extensive and expensive simulations.Alongside the evolution of nanophotonics, various methods for expediting the design of photonic structures have been developed. Gradient-based adjoint methods, such as topology optimization, are a class of widely applied approaches for Molecules composed of atoms exhibit properties not inherent to their constituent atoms. Similarly, metamolecules consisting of multiple meta-atoms possess emerging features that the meta-atoms themselves do not possess. Metasurfaces composed of metamolecules with spatially variant building blocks, such as gradient metasurfaces, are drawing substantial attention due to their unconventional controllability of the amplitude, phase, and frequency of light. However, the intricate mechanisms and the large degrees of freedom of the multielement systems impede an effective strategy for the design and optimization of metamolecules. Here, a hybrid artificial-i...

show abstract