Reproduce “Large-Area Fabrication-Aware Computational Diffractive Optics” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kaixuan Wei, Hector A. Jimenez-Romero, Hadi Amata, Jipeng Sun, Qiang Fu, Felix Heide, Wolfgang Heidrich
Researchers at KAUST and Princeton developed a fabrication-aware design pipeline for diffractive optical elements, integrating a super-resolved neural lithography model and a tensor-parallel computing framework. This approach enables the accurate design and large-scale simulation of DOEs, yielding physical prototypes that closely match simulated performance, such as 4K × 6K pixel holograms and beam splitters with 53% higher experimental spot intensity.
Key insights
- Explicitly modeling fabrication-induced distortions within the design optimization process significantly bridges the simulation-to-reality gap, leading to more faithful physical prototypes.
- Data-driven neural networks can effectively learn and predict the complex, non-linear physical effects of micro/nano-fabrication, outperforming traditional physical models in accuracy.
- Distributed tensor-parallel computing is essential for scaling diffractive optics design to centimeter-scale devices with sub-micron features, overcoming the memory and computational bottlenecks of single-GPU setups.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.