Artificial intelligence is transforming the design of optical metasurfaces, overcoming key challenges that have hindered their development from unit-cell optimization to full system integration according to a new review article. Optical metasurfaces, known for their ultra-thin and lightweight properties, are crucial for miniaturizing optical systems, but their complex design process has presented significant obstacles that AI is now addressing at each design stage.
At the unit-cell level, AI-driven surrogate modeling accelerates electromagnetic response prediction while inverse design frameworks explore complex solution spaces. Robust design methods enhance stability against manufacturing variations, which has been a persistent challenge in metasurface fabrication. For metasurface optimization, AI methods like graph neural networks model non-local interactions between densely packed meta-atoms according to Professor Xin Jin from Tsinghua University, who led the review. Multi-task learning resolves conflicting performance objectives, and reinforcement learning enables real-time dynamic control in these advanced optical systems.
At the system level, AI provides a unified differentiable framework that integrates structural design, physical propagation models, and task-specific loss functions. This end-to-end optimization directly links nanostructure design to final application goals, overcoming incompatibility between metasurface design and backend algorithms that has previously limited practical implementation. The integration of AI allows designers to optimize metasurfaces for specific functions rather than treating them as isolated components, enabling more efficient development of advanced applications including compact imaging systems, augmented and virtual reality displays, advanced LiDAR, and computational imaging systems.
The review identifies several future research directions, including developing AI methods more deeply integrated with electromagnetic theory, creating unified architectures for multi-scale design, and advancing adaptive photonic platforms. These developments could further accelerate the adoption of metasurface technology across multiple industries. The original research is available at https://doi.org/10.1016/j.iopt.2025.100004 with funding from multiple sources including the Shenzhen Science and Technology Program, Natural Science Foundation of China, and the Major Key Project of PCL. This research represents a significant step toward making metasurface technology more accessible and practical for real-world applications, potentially transforming fields from consumer electronics to scientific instrumentation through more compact, efficient optical systems that were previously limited by design constraints.
