BootAug: Boosting Text Augmentation via Hybrid Instance Filtering Framework

arXiv:2210.02941v2 Announce Type: replace
Abstract: 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 $\approx 2\%$ in aspect-based sentiment classification). To address this …

abstract arxiv augmentation boosting cs.cl data datasets few-shot filtering focus framework generate however hybrid instance instances language language processing natural natural language natural language processing processing public research text type via