Inverse Design of Grating Couplers Using the Policy Gradient Method from Reinforcement Learning
arXiv:2107.00088 · doi:10.1515/nanoph-2021-0332
Abstract
We present a proof-of-concept technique for the inverse design of electromagnetic devices motivated by the policy gradient method in reinforcement learning, named PHORCED (PHotonic Optimization using REINFORCE Criteria for Enhanced Design). This technique uses a probabilistic generative neural network interfaced with an electromagnetic solver to assist in the design of photonic devices, such as grating couplers. We show 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. As a further example of the benefits of this method, we implement transfer learning with PHORCED, demonstrating that a neural network trained to optimize 8 grating couplers can then be re-trained on grating couplers with alternate scattering angles while requiring >10 fewer simulations than control cases.
arXiv commentMain Document: 14 pages, 4 figures; Supplementary Document: 6 pages, 2 figures
References in corpus (26)
- Deep Learning with Coherent Nanophotonic Circuits
- Outlook for inverse design in nanophotonics
- Photonics for artificial intelligence and neuromorphic computing
- 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
- Inverse design and demonstration of a compact and broadband on-chip wavelength demultiplexer
- Evolution Strategies as a Scalable Alternative to Reinforcement Learning
- Physics-informed neural networks for inverse problems in nano-optics and metamaterials
- Deep neural networks for the evaluation and design of photonic devices
- Quantum transport simulations in a programmable nanophotonic processor
- Global optimization of dielectric metasurfaces using a physics-driven neural network
- Wave Physics as an Analog Recurrent Neural Network
- Adjoint method and inverse design for nonlinear nanophotonic devices
- Inverse design of photonic crystals through automatic differentiation
- Inverse Design of Near Unity Efficiency Perfectly Vertical Grating Couplers
- Simulator-based training of generative models for the inverse design of metasurfaces
- Mapping the global design space of nanophotonic components using machine learning pattern recognition
- Fully-automated optimization of grating couplers
- Inverse designed metalenses with extended depth of focus
- Active learning of deep surrogates for PDEs: Application to metasurface design
- Designing Silicon Photonic Devices using Artificial Neural Networks
- Leveraging Continuous Material Averaging for Inverse Electromagnetic Design
- Multi-objective and categorical global optimization of photonic structures based on ResNet generative neural networks
- Design of Compact and Efficient Silicon Photonic Micro Antennas with Perfectly Vertical Emission
- Inverse Design of Compact Multimode Cavity Couplers
- Improving Evolutionary Strategies with Generative Neural Networks
- Neural Generative Models for Global Optimization with Gradients