Data-driven acceleration of photonic simulations

@article{Trivedi2019DatadrivenAO,
  title={Data-driven acceleration of photonic simulations},
  author={Rahul Trivedi and Logan Su and Jesse Lu and Martina Schubert and Jelena Vu{\vc}kovi{\'c}},
  journal={Scientific Reports},
  year={2019},
  volume={9}
}
Designing modern photonic devices often involves traversing a large parameter space via an optimization procedure, gradient based or otherwise, and typically results in the designer performing electromagnetic simulations of a large number of correlated devices. In this paper, we investigate the possibility of accelerating electromagnetic simulations using the data collected from such correlated simulations. In particular, we present an approach to accelerate the Generalized Minimal Residual… 

Deep neural networks for the evaluation and design of photonic devices

TLDR
This Review discusses how deep neural networks can serve as surrogate electromagnetic solvers, inverse modelling tools and global device optimizers, and how deep generative networks can learn geometric features in device distributions and even be configured to serve as robust global optimizers.

Simulation acceleration for transmittance of electromagnetic waves in 2D slit arrays using deep learning

TLDR
A deep learning approach is introduced that accelerates a simulator solving frequency-domain Maxwell equations and achieves high accuracy while predicting transmittance per wavelength in 2D slit arrays under certain conditions to achieve 160,000 times faster results than the simulator.

Photonic Inverse Design with Neural Networks: The Case of Invisibility in the Visible

TLDR
A fully connected neural network is proposed to address the problem of invisible nanoparticles by learning the dynamics of visible-light interaction with low-scattering multilayered nanospheres, and can be generalized to approximate Maxwell interactions by simulating the electromagnetic response of more complicated optical configurations.

MaxwellNet: Physics-driven deep neural network training based on Maxwell’s equations

TLDR
This paper presents a novel scheme to train a DNN that solves Maxwell's equations speedily and accurately without relying on other computational electromagnetic solvers, and exploits the speed of this network in a novel inverse design scheme to design a micro-lens that maximizes the desired merit function.

WaveY-Net: physics-augmented deep-learning for high-speed electromagnetic simulation and optimization

We introduce WaveY-Net, a hybrid data- and physics-augmented convolutional neural network that can predict electromagnetic field distributions with ultra fast speeds and high accuracy for entire

Deep learning in nano-photonics: inverse design and beyond

TLDR
A critical review on the capabilities of deep learning for inverse design and the progress which has been made so far, and classify the different deep learning-based inverse design approaches at a higher level as well as by the context of their respective applications.

Interfacing photonics with artificial intelligence: an innovative design strategy for photonic structures and devices based on artificial neural networks

TLDR
This review will introduce several commonly used neural networks and highlight their applications in the design process of various optical structures and devices, particularly those in recent experimental works and comment on the future directions to inspire researchers from different disciplines to collectively advance this emerging research field.

Inverse design enables large-scale high-performance meta-optics reshaping virtual reality

Meta-optics has achieved major breakthroughs in the past decade; however, conventional forward design faces challenges as functionality complexity and device size scale up. Inverse design aims at

Inverse design of grating couplers using the policy gradient method from reinforcement learning

TLDR
It is shown that PHORCED obtains better performing grating coupler designs than local gradient-based inverse design via the adjoint method, while potentially providing faster convergence over competing state-of-the-art generative methods.

Machine learning accelerated particle-in-cell plasma simulations

TLDR
This work investigates how amortized solvers can be incorporated with PIC methods for simulations of plasmas and finds that this approach reduces the average number of required solver iterations by about 25% when simulating electron plasma oscillations.

References

SHOWING 1-10 OF 32 REFERENCES

Effective index approximations of photonic crystal slabs: a 2-to-1-D assessment

The optical properties of slab-like photonic crystals are often discussed on the basis of effective index (EI) approximations, where a 2-D effective refractive index profile replaces the actual 3-D

Large‐scale topology optimization using preconditioned Krylov subspace methods with recycling

TLDR
It is shown that a proper rescaling of the linear systems reduces the huge condition numbers that typically occur in topology optimization to roughly those arising for a problem with constant density.

Recycling Krylov Subspaces for Sequences of Linear Systems

TLDR
This work proposes and analyzes two methods that significantly reduce the total number of matrix-vector products required to solve all systems and can reduce the cost of solving subsequent systems in the sequence by recycling selected subspaces generated for previous systems.

Solving high-dimensional partial differential equations using deep learning

TLDR
A deep learning-based approach that can handle general high-dimensional parabolic PDEs using backward stochastic differential equations and the gradient of the unknown solution is approximated by neural networks, very much in the spirit of deep reinforcement learning with the gradient acting as the policy function.

Training Deep Neural Networks for the Inverse Design of Nanophotonic Structures

We demonstrate a tandem neural network architecture that tolerates inconsistent training instances in inverse design of nanophotonic devices. It provides a way to train large neural networks for the

Adam: A Method for Stochastic Optimization

TLDR
This work introduces Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments, and provides a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework.

Inverse design and implementation of a wavelength demultiplexing grating coupler

TLDR
A general inverse design algorithm for designing micro- and nano-photonic devices, where the user only specifies the desired device performance, is developed.

A Simplified and Flexible Variant of GCROT for Solving Nonsymmetric Linear Systems

TLDR
This work develops a flexible variant of the Krylov linear solver GCROT (generalized conjugate residual with inner orthogonalization and outer truncation) that is robust and efficient compared with several popular truncated Krylov subspace methods.

Inverse design and demonstration of a compact and broadband on-chip wavelength demultiplexer

An on-chip integrated wavelength demultiplexer designed using an inverse computational algorithm is experimentally demonstrated. 1,300 and 1,550 nm wavelength light is sorted in a device area of just

TensorFlow: learning functions at scale

TLDR
This talk describes Tensor Flow and outlines some of its applications, and discusses the question of what TensorFlow and deep learning may have to do with functional programming.