Lilian Weng • 4/16/2022

Learning with not Enough Data Part 3: Data Generation

This technical article details strategies for generating synthetic data when real-world data is scarce. It covers data augmentation techniques for images and text, including task-specific methods like AutoAugment and RandAugment, as well as using large pretrained models for data creation. It is part three of a series on learning with insufficient data.

0 comments

#Machine Learning #image processing #Data Augmentation