Researchers Train AI to Design Optical Surfaces Accounting for Real‑World Imperfections

Scientists have developed an artificial‑intelligence system that designs optical surfaces. The AI model incorporates real‑world imperfections typically found in lenses and mirrors. Training data includes measurements of manufacturing defects and material variations. The approach enables the AI to propose designs that are robust to such flaws. Results suggest improved performance over conventional design methods. The work was reported through EurekAlert! as a research breakthrough. Potential applications span imaging systems, telescopes, and consumer optics. Further testing will assess the AI‑generated designs in practical manufacturing settings.