Good-Enough Example Extrapolation
This addresses data imbalance in text classification, but it is incremental as it builds on existing hidden-space augmentation methods.
The paper tackles the problem of data augmentation for text classification by proposing a lightweight method called good-enough example extrapolation (GE3), which extrapolates hidden space distributions from one class to another, and shows that it improves performance more than upsampling and other methods on three datasets in imbalanced scenarios.
This paper asks whether extrapolating the hidden space distribution of text examples from one class onto another is a valid inductive bias for data augmentation. To operationalize this question, I propose a simple data augmentation protocol called "good-enough example extrapolation" (GE3). GE3 is lightweight and has no hyperparameters. Applied to three text classification datasets for various data imbalance scenarios, GE3 improves performance more than upsampling and other hidden-space data augmentation methods.