Table 1 This table shows the device types (see Fig. 1a) their area, the number of devices of each time across all 10 wafers, the percentage of devices that passed both forward and reverse criteria of each type, and the accuracy of the model using a two layer neural network model constructed by using halve of the devices of the size of interest as the validation test and the remaining data as the training data set.

From: Using machine learning with optical profilometry for GaN wafer screening