Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

BOSON$^{-1}$: Understanding and Enabling Physically-Robust Photonic Inverse Design with Adaptive Variation-Aware Subspace Optimization

Pingchuan Ma, Zhengqi Gao, Amir Begovic, Meng Zhang, Haoyu Yang, Haoxing Ren, Zhaoran Rena Huang, Duane Boning, Jiaqi Gu

Abstract

Nanophotonic device design aims to optimize photonic structures to meet specific requirements across various applications. Inverse design has unlocked non-intuitive, high-dimensional design spaces, enabling the discovery of high-performance devices beyond heuristic or analytic methods. The adjoint method, which calculates gradients for all variables using just two simulations, enables efficient navigation of this complex space. However, many inverse-designed structures, while numerically plausible, are difficult to fabricate and sensitive to variations, limiting their practical use. The discrete nature with numerous local-optimal structures also pose significant optimization challenges, often causing gradient-based methods to converge on suboptimal designs. In this work, we formulate inverse design as a fabrication-restricted, discrete, probabilistic optimization problem and introduce BOSON-1, an end-to-end, variation-aware subspace optimization framework to address the challenges of manufacturability, robustness, and optimizability. To overcome optimization difficulty, we propose dense target-enhanced gradient flows to mitigate misleading local optima and introduce a conditional subspace optimization strategy to create high-dimensional tunnels to escape local optima. Furthermore, we significantly reduce the runtime associated with optimizing across exponential variation samples through an adaptive sampling-based robust optimization, ensuring both efficiency and variation robustness. On three representative photonic device benchmarks, our proposed inverse design methodology BOSON^-1 delivers fabricable structures and achieves the best convergence and performance under realistic variations, outperforming prior arts with 74.3% post-fabrication performance. We open-source our codes at https://github.com/ScopeX-ASU/BOSON.

BOSON$^{-1}$: Understanding and Enabling Physically-Robust Photonic Inverse Design with Adaptive Variation-Aware Subspace Optimization

Abstract

Nanophotonic device design aims to optimize photonic structures to meet specific requirements across various applications. Inverse design has unlocked non-intuitive, high-dimensional design spaces, enabling the discovery of high-performance devices beyond heuristic or analytic methods. The adjoint method, which calculates gradients for all variables using just two simulations, enables efficient navigation of this complex space. However, many inverse-designed structures, while numerically plausible, are difficult to fabricate and sensitive to variations, limiting their practical use. The discrete nature with numerous local-optimal structures also pose significant optimization challenges, often causing gradient-based methods to converge on suboptimal designs. In this work, we formulate inverse design as a fabrication-restricted, discrete, probabilistic optimization problem and introduce BOSON-1, an end-to-end, variation-aware subspace optimization framework to address the challenges of manufacturability, robustness, and optimizability. To overcome optimization difficulty, we propose dense target-enhanced gradient flows to mitigate misleading local optima and introduce a conditional subspace optimization strategy to create high-dimensional tunnels to escape local optima. Furthermore, we significantly reduce the runtime associated with optimizing across exponential variation samples through an adaptive sampling-based robust optimization, ensuring both efficiency and variation robustness. On three representative photonic device benchmarks, our proposed inverse design methodology BOSON^-1 delivers fabricable structures and achieves the best convergence and performance under realistic variations, outperforming prior arts with 74.3% post-fabrication performance. We open-source our codes at https://github.com/ScopeX-ASU/BOSON.

Paper Structure

This paper contains 27 sections, 3 equations, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminary
  3. Fabrication and Operation Variations
  4. Prior Fabrication-Robust Photonic Inverse Design Methods
  5. Robust Photonic Inverse Design with Adaptive Variation-Aware Subspace Optimization
  6. Problem Formulation
  7. Understanding the Fabricability Constraint and Optimization Difficulty of Photonic Inverse Design
  8. Fabrication Constraints
  9. Optimization Difficulty
  10. Fabrication-Restricted Subspace Optimization Toward Guaranteed Manufacturability
  11. Gradient-Enhanced Subspace Adjoint Optimization Toward Better Optimality
  12. Gradient-Enhanced Objective Landscape Reshaping via Auxiliary Constraints
  13. Subspace Local Optima Escaping via Relaxed High-Dimensional Tunneling
  14. Optical Path Concentrated Initialization
  15. Adaptive Sampling Strategy for Efficiency and Robustness
  16. ...and 12 more sections

Figures (6)

  • Figure 1: Inverse design often yields non-fabricable devices. Optimization difficulty leads to suboptimal designs.
  • Figure 2: (a) Lithography and etching during fabrication restricts manufacturable patterns in the subspace. (b) Fabrication variations (defocusing and etching) and operation variations cause performance deviation.
  • Figure 3: Our proposed BOSON$^{-1}$ framework enables efficient, robust optimization with better convergence and optimality.
  • Figure 4: High-dimensional tunnel for local minima escaping.
  • Figure 5: Fabrication-aware optimization trajectories of optical isolator with forward/backward transmission, radiation, and reflection. No variation is added. (a) Proposed method (light-concentrated initialization, dense objectives). (b) Light-concentrated initialization with contrast as the single sparse objective. (c) Random initialization with contrast as the single sparse objective.
  • ...and 1 more figures