FlipDA: Effective and Robust Data Augmentation for Few-Shot Learning
This addresses the problem of robust data augmentation for few-shot learning in NLP, offering a novel approach that is effective across various tasks, though it is incremental in improving upon existing augmentation techniques.
The paper tackled the challenge of data augmentation for few-shot natural language understanding with strong pretrained models, where existing methods often yield marginal gains or degrade performance. The proposed FlipDA method, which generates label-flipped data using a generative model and classifier, substantially improves many tasks without negatively affecting others.
Most previous methods for text data augmentation are limited to simple tasks and weak baselines. We explore data augmentation on hard tasks (i.e., few-shot natural language understanding) and strong baselines (i.e., pretrained models with over one billion parameters). Under this setting, we reproduced a large number of previous augmentation methods and found that these methods bring marginal gains at best and sometimes degrade the performance much. To address this challenge, we propose a novel data augmentation method FlipDA that jointly uses a generative model and a classifier to generate label-flipped data. Central to the idea of FlipDA is the discovery that generating label-flipped data is more crucial to the performance than generating label-preserved data. Experiments show that FlipDA achieves a good tradeoff between effectiveness and robustness -- it substantially improves many tasks while not negatively affecting the others.