Boosting Text Augmentation via Hybrid Instance Filtering Framework
Heng Yang, Ke Li · 2023
Text augmentation is an effective technique for addressing the problem of insufficient data in natural language processing.However, existing text augmentation methods tend to focus on few-shot scenarios and usually perform poorly on large public datasets.Our research indicates that existing augmentation methods often generate instances with shifted feature spaces, which leads to a drop in performance on the augmented data (for example, EDA generally loses ≈ 2% in aspect-based sentiment classification).To address this problem, we propose a hybrid instance-filtering framework (BO O S TAU G) based on pre-trained language models that can maintain a similar feature space with natural datasets.BO O S TAU G is transferable to existing text augmentation methods (such as synonym substitution and back translation) and significantly improves the augmentation performance by ≈ 2 -3% in classification accuracy.Our experimental results on three classification tasks and nine public datasets show that BO O S TAU G addresses the performance drop problem and outperforms state-ofthe-art text augmentation methods.Additionally, we release the code to help improve existing augmentation methods on large datasets.