supplementary material

hyperparameters of the Convolutional Neural Network(CNN) model, a numerical figure of the loss function versus the number of iterations during CNN training on regular-patterned surfaces, results of 5-fold cross-validation, data from the image robustness test, examples of regular and irregular hydroxyl functionalization patterns on GO surfaces, prediction system diagrams (including the prediction process for direct prediction on irregular-patterned surfaces), and a detailed data table on the fraction of high-density regions in the density distribution.