Plasmonic Metamaterial Absorbers Design Based on XGBoost and LightGBM Algorithms

The emergence of metamaterials has brought a revolutionary way to manipulate the behavior of light on the nanoscale. However, there are still many problems in design process, such as time-consuming and many-to-one mapping. Here, we demonstrate the forward and inverse design of plasmonic metamaterial absorbers based on Light gradient boosting machine (LightGBM) and Extreme Gradient Boosting (XGBoost). The inverse framework can use the input reflection value to design the metamaterial parameter structure. The experimental results show that XGBoost has better performance in forward and inverse design (Forward-R-2: 0.956; Inverse-R-2: 0.967). The framework is suitable for designing metamaterials on demand, and it can be used in zoom imaging, metamaterial absorbers, metamaterial filters, and other fields.