You do data augmentation in deep learning (in the context of images dataset for example) for solve unbalance classes problem, so you augmented the number of images in the minoritary class. Other reason is, if you have low amount of data. A very popular technique consist in rotate or fliped the images in this way you create new images for the dataset.

Dear Doctor

"Data augmentation is a technique of artificially increasing the training set by creating modified copies of a dataset using existing data. It includes making minor changes to the dataset or using deep learning to generate new data points."