Physics-guided and fabrication-aware inverse design of photonic devices using diffusion models

2025-04-23

AI Summary

Photonic device inverse design faces challenges including an enormous geometric search space, complex fabrication constraints, and high computational cost of electromagnetic simulations. To address these, we propose AdjointDiffusion—a novel framework that intrinsically integrates adjoint-sensitivity gradients into the denoising process of diffusion models, enabling physics-informed and fabrication-aware joint optimization. Unlike conventional approaches, AdjointDiffusion directly synthesizes high-performance, manufacturable structures in a single sampling pass—eliminating the need for post-hoc binarization or morphological filtering. Evaluated on bent waveguide routing and CMOS image sensor color-routing tasks, it achieves state-of-the-art optical performance and yield while requiring only ~200 full-wave simulations—reducing simulation count by 3–4 orders of magnitude compared to pure deep learning methods—and outperforming conventional optimizers (e.g., MMA, SLSQP) in both convergence speed and solution quality.

Technology Category

• Search and Optimization: Adversarial Search
• Computer Vision: Diffusion Models for Vision
• Machine Learning: Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Ad search and search for Web retail
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Designing free-form photonic devices is fundamentally challenging due to the vast number of possible geometries and the complex requirements of fabrication constraints. Traditional inverse-design approaches--whether driven by human intuition, global optimization, or adjoint-based gradient methods--often involve intricate binarization and filtering steps, while recent deep learning strategies demand prohibitively large numbers of simulations (10^5 to 10^6). To overcome these limitations, we present AdjointDiffusion, a physics-guided framework that integrates adjoint sensitivity gradients into the sampling process of diffusion models. AdjointDiffusion begins by training a diffusion network on a synthetic, fabrication-aware dataset of binary masks. During inference, we compute the adjoint gradient of a candidate structure and inject this physics-based guidance at each denoising step, steering the generative process toward high figure-of-merit (FoM) solutions without additional post-processing. We demonstrate our method on two canonical photonic design problems--a bent waveguide and a CMOS image sensor color router--and show that our method consistently outperforms state-of-the-art nonlinear optimizers (such as MMA and SLSQP) in both efficiency and manufacturability, while using orders of magnitude fewer simulations (approximately 2 x 10^2) than pure deep learning approaches (approximately 10^5 to 10^6). By eliminating complex binarization schedules and minimizing simulation overhead, AdjointDiffusion offers a streamlined, simulation-efficient, and fabrication-aware pipeline for next-generation photonic device design. Our open-source implementation is available at https://github.com/dongjin-seo2020/AdjointDiffusion.

Problem

• Overcoming fabrication constraints in photonic device design
• Reducing simulation costs in inverse-design approaches
• Integrating physics guidance into diffusion model sampling

Innovation

• Physics-guided diffusion models for photonic design
• Adjoint gradient integration in denoising steps
• Fabrication-aware dataset training for binary masks