Algorithm 1 Active learning using surrogate model as the oracle

Input: Define the parameter space. N×N metasurface array with various combinations of elements 0 and 1. Perform: CST simulation to calculate the RCS values over desired frequency range for considered metasurface array. Collect: Collect training dataset. Construct: Build the appropriate Gaussian process regression model. while $budget\ne 0$ do    Find: The metasurface array combination with minimum $RCSpredictions$ in the frequency range in the test set.    Compute: Train Gaussian process regressor’s prediction to calculate the RCS value of the metasurface array from the above step.    Inclusion: Include the above combination of the metasurface array and its corresponding RCS values in the training dataset.    Retrain: Train the model again using the above informative sample and its respective output.    $budget\leftarrow budget-1$ end while Optimal design: Select the candidate metasurface array contributing to minimum RCS in the desired frequency range.