IndiText Boost: Text Augmentation for Low Resource India Languages
Onkar Litake, Niraj Yagnik, Shreyas Labhsetwar
TL;DR
This paper tackles data scarcity in text classification for six low-resource Indian languages by systematically evaluating eight augmentation techniques, including Easy Data Augmentation (EDA), back-translation, paraphrasing, and LLM-based methods. Using a uniform pipeline, the authors generate one augmented example per original sentence and fine-tune a multilingual BERT, comparing results across binary and multiclass tasks. Across languages, most augmentation methods improve performance, with EDA often delivering the strongest gains, while some GPT-based approaches are hampered by API limits and shorter augmented texts. The findings highlight practical augmentation strategies for low-resource NLP and point to limitations in language coverage and compute that future work should address to broaden applicability.
Abstract
Text Augmentation is an important task for low-resource languages. It helps deal with the problem of data scarcity. A data augmentation strategy is used to deal with the problem of data scarcity. Through the years, much work has been done on data augmentation for the English language. In contrast, very less work has been done on Indian languages. This is contrary to the fact that data augmentation is used to deal with data scarcity. In this work, we focus on implementing techniques like Easy Data Augmentation, Back Translation, Paraphrasing, Text Generation using LLMs, and Text Expansion using LLMs for text classification on different languages. We focus on 6 Indian languages namely: Sindhi, Marathi, Hindi, Gujarati, Telugu, and Sanskrit. According to our knowledge, no such work exists for text augmentation on Indian languages. We carry out binary as well as multi-class text classification to make our results more comparable. We get surprising results as basic data augmentation techniques surpass LLMs.