Abstract: Neural networks (NNs) have emerged as powerful tools for accelerating the design of frequency-selective surfaces (FSSs) due to their fast prediction of electromagnetic (EM) responses.