nmT5 -- Is parallel data still relevant for pre-training massively multilingual language models?
This addresses the efficiency of data usage for multilingual NLP, with incremental insights on model scaling.
The paper investigated whether parallel data improves pre-training for massively multilingual language models like mT5, finding that it boosts performance on multilingual tasks but gains diminish with larger models, though benefits persist in low-data scenarios.
Recently, mT5 - a massively multilingual version of T5 - leveraged a unified text-to-text format to attain state-of-the-art results on a wide variety of multilingual NLP tasks. In this paper, we investigate the impact of incorporating parallel data into mT5 pre-training. We find that multi-tasking language modeling with objectives such as machine translation during pre-training is a straightforward way to improve performance on downstream multilingual and cross-lingual tasks. However, the gains start to diminish as the model capacity increases, suggesting that parallel data might not be as essential for larger models. At the same time, even at larger model sizes, we find that pre-training with parallel data still provides benefits in the limited labelled data regime.