Augmenting Data with Mixup for Sentence Classification: An Empirical Study

TL;DR

This paper investigates the application of Mixup data augmentation techniques to sentence classification in NLP, proposing two embedding interpolation strategies and demonstrating their effectiveness across multiple datasets and models.

Contribution

It introduces two novel Mixup-based augmentation methods for NLP sentence classification and empirically evaluates their effectiveness, filling a gap in applying Mixup to natural language tasks.

Findings

Interpolation strategies improve accuracy for CNN and LSTM models

Mixup augmentation is effective across multiple benchmark datasets

Proposed methods are domain independent

Abstract

Mixup, a recent proposed data augmentation method through linearly interpolating inputs and modeling targets of random samples, has demonstrated its capability of significantly improving the predictive accuracy of the state-of-the-art networks for image classification. However, how this technique can be applied to and what is its effectiveness on natural language processing (NLP) tasks have not been investigated. In this paper, we propose two strategies for the adaption of Mixup on sentence classification: one performs interpolation on word embeddings and another on sentence embeddings. We conduct experiments to evaluate our methods using several benchmark datasets. Our studies show that such interpolation strategies serve as an effective, domain independent data augmentation approach for sentence classification, and can result in significant accuracy improvement for both CNN and LSTM models.