{
  "id": 156533,
  "title": "Yield-aware generative inverse design of anti-reflection coatings via optimal-transport flow matching with a conditional-value-at-risk objective",
  "url": "https://urgent.news/2026/08/05/yield-aware-generative-inverse-design-of-anti-reflection-coatings-via",
  "topic": "world",
  "section": "World",
  "published": "2026-08-05T00:00:00.000Z",
  "source": {
    "name": "Scientific Reports",
    "slug": "scientific-reports",
    "url": "https://www.nature.com/articles/s41598-026-65513-5"
  },
  "original_language": "en",
  "account": "In a recent study, researchers have developed a new approach to designing anti-reflection (AR) coatings that is more robust to manufacturing errors. Traditionally, AR coating design would first find an optimal stack and then assess its tolerance to deposition errors. This new method, however, incorporates robustness directly into the design objective.\n\nRather than minimizing the mean reflectance, the researchers minimized the conditional value-at-risk (CVaR) of the band-mean reflectance under the distribution of deposition errors. This yield-aware formulation directs the optimizer towards the worst-case scenarios rather than relying on an idealized nominal spectrum. The researchers combined this yield-aware approach with an Optimal-Transport Conditional Flow Matching (OT-CFM) generator, which operates under a manufacturability constraint enforced by a smooth box reparameterization. They also used a differentiable transfer-matrix forward model validated against the analytic quarter-wave result for machine precision.\n\nTo test the effectiveness of their approach, the researchers evaluated a six-layer MgF2/SiO2 stack on 10^5 held-out deposition realizations. The CVaR-robust design reduced the 99th-percentile band-mean reflectance from 1.825 to 1.810 pp. While this improvement in reflectance is small in absolute terms, the researchers report it alongside the sampling uncertainty to establish its significance.\n\nThe advantage of the CVaR-robust design becomes evident when the deposition error is greater than 4 nm, and it plateaus above 12 nm. However, a material-choice ablation showed that introducing high-index TiO2 degrades AR performance within the tested design space. The authors acknowledge the helpful feedback from anonymous reviewers and Dr. Mostafa Salahshoor, as well as the assistance of an AI assistant for code debugging and language editing. The study is licensed under a Creative Commons Attribution 4.0 International License.",
  "summary": "Scientific Reports, Published online: 05 August 2026; doi:10.1038/s41598-026-65513-5 Yield-aware generative inverse design of anti-reflection coatings via optimal-transport flow matching with a conditional-value-at-risk objective",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}