It offers unique DA challenges as there are typically many correlated classification decisions.

University of Michigan - Ann Arbor

1. Dissonance between empirical novelties and theoretical narrative
2. Minimal benefit for pretrained models on indomain data
3. Multimodal challenges
4. Span-based tasks
5. Working in specialized domains
6. Working with low-resource languages
7. More vision-inspired techniques
8. Self-supervised learning
9. Offline versus online data augmentation
10. Lack of unification
11. Good data augmentation practices

Challenges & Future Directions of DA in NLP

https://arxiv-download.xixiaoyao.cn/pdf/2105.03075.pdf

A Survey of Data Augmentation Approaches for NLP

Effective DA methods for multiple modalities have been challenging. Many works focus on augmenting a single modality or multiple ones separately.

Multimodal Challenges of DA in NLP

Why Span-based Tasks are Important for DA in NLP 

DA techniques successful in the high resource scenario may not be effective for low resource languages that are of a different language family or very distinctive in linguistic and typological terms. 

Learn Before

Related