The scientists developed a method to correct and enhance spectroscopic data by using machine learning and modeling techniques. They focused on removing and modeling variations in the data that come from having multiple samples. By applying this method to different types of spectra from yeasts, fungi, and bacteria, they found that correcting variations in the data improved the accuracy of identifying different species. They also discovered that adding variations to the data through a process called augmentation helped deep learning algorithms perform better at classifying the samples.