Exploring the Role of Transliteration in In-Context Learning for Low-resource Languages Written in Non-Latin Scripts
This addresses the challenge of adapting LLMs to low-resource languages for NLP applications, but it is incremental as it builds on existing transliteration methods.
The paper tackled the problem of low performance of large language models in low-resource languages with non-Latin scripts by exploring transliteration in in-context learning, finding that transliteration improved performance by up to 25% in sequential labeling tasks.
Decoder-only large language models (LLMs) excel in high-resource languages across various tasks through few-shot or even zero-shot in-context learning (ICL). However, their performance often does not transfer well to low-resource languages, especially those written in non-Latin scripts. Inspired by recent work that leverages transliteration in encoder-only models, we investigate whether transliteration is also effective in improving LLMs' performance for low-resource languages written in non-Latin scripts. To this end, we propose three prompt templates, where the target-language text is represented in (1) its original script, (2) Latin script, or (3) both. We apply these methods to several representative LLMs of different sizes on various tasks including text classification and sequential labeling. Our findings show that the effectiveness of transliteration varies by task type and model size. For instance, all models benefit from transliterations for sequential labeling (with increases of up to 25%).