Systems and methods for discovering defects on a wafer are provided. One method includes detecting defects on a wafer by applying a threshold to output generated by a detector in a first scan of the wafer and determining values for features of the detected defects. The method also includes automatically ranking the features, identifying feature cut-lines to group the defect into bins, and, for each of the bins, determining one or more parameters that if applied to the values for the features of the defects in each of the bins will result in a predetermined number of the defects in each of the bins. The method also includes applying the one or more determined parameters to the output generated by the detector in a second scan of the wafer to generate a defect population that has a predetermined defect count and is diversified in the values for the features.
