EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks

Jason Wei, Kai Zou
1/31/2019
cs.CL

Official code: jasonwei20/eda_nlp (1,651 stars, last commit 2023-03-19).

Abstract

We present EDA: easy data augmentation techniques for boosting performance on text classification tasks. EDA consists of four simple but powerful operations: synonym replacement, random insertion, random swap, and random deletion. On five text classification tasks, we show that EDA improves performance for both convolutional and recurrent neural networks. EDA demonstrates particularly strong results for smaller datasets; on average, across five datasets, training with EDA while using only 50% of the available training set achieved the same accuracy as normal training with all available data. We also performed extensive ablation studies and suggest parameters for practical use.

View on arXivView PDF

Code Implementations(1)

jasonwei20/eda_nlpOfficial100%

Data augmentation for NLP, presented at EMNLP 2019

1,651312PythonDec 27, 20183 years ago
classificationcnndata-augmentationembeddingsnlp+6 more

Similar Papers

Textual Augmentation Techniques Applied to Low Resource Machine Translation: Case of Swahili

85% match2023

Distributional Data Augmentation Methods for Low Resource Language

85% match2023

Exploring Data Augmentation Methods on Social Media Corpora

83% match2023

Performance of Data Augmentation Methods for Brazilian Portuguese Text Classification

83% match2023

Adversarial Word Dilution as Text Data Augmentation in Low-Resource Regime

82% match2023

Bag of Tricks for Efficient Text Classification

82% match1 repo26.5k2016

Cite this paper

@article{wei2019eda,
  title  = {EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks},
  author = {Jason Wei and Kai Zou},
  year   = {2019},
  eprint = {1901.11196},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1901.11196}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1901.11196.svg)](https://codeofpaper.com/paper/1901.11196)

Discussion