Physics-guided and fabrication-aware inverse design of photonic devices using diffusion models
Dongjin Seo, Soobin Um, Sangbin Lee, Jong Chul Ye, Haejun Chung · 2025
Read Physics-guided and fabrication-aware inverse design of photonic devices using diffusion models by Dongjin Seo, Soobin Um, Sangbin Lee, Jong Chul Ye, Haejun Chung (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-17077.txt.
Description
Dongjin Seo and four co-authors present an AI method for designing light devices in a 2025 paper. Free-form designs for lenses and light routers live in a huge space of shapes, and factory rules narrow it further. Older design methods are fiddly, and pure machine learning ones need hundreds of thousands of simulations.
Their framework first trains a generative model on synthetic, factory-friendly shapes. Then, while the model drafts a design, physics-based gradients from a sensitivity method steer each step toward better performance. No cleanup steps are needed afterward. On two standard problems, a bent light guide and a color-splitting sensor part, the method beat top classic optimizers on both quality and ease of manufacture. It also used only a few hundred simulations, far fewer than pure learning approaches. The blend of physics with generation is the paper's core idea.
Catalog record
| CREATOR | Dongjin Seo, Soobin Um, Sangbin Lee, Jong Chul Ye, Haejun Chung |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-17077