Training Bilingual LMs with Data Constraints in the Targeted Language
This addresses the challenge of improving language model performance for low-resource languages, though it is incremental as it builds on existing translation and data upsampling methods.
The paper tackles the problem of training bilingual language models when high-quality pretraining data is scarce in the target language, by leveraging data from an auxiliary language like English, and finds that stronger auxiliary datasets lead to performance gains without model modifications, especially for closely related languages.
Large language models are trained on massive scrapes of the web, as required by current scaling laws. Most progress is made for English, given its abundance of high-quality pretraining data. For most other languages, however, such high quality pretraining data is unavailable. In this work, we study how to boost pretrained model performance in a target language with insufficient pretraining data for training a high performing language model, by enlisting data from an auxiliary language for which high quality data is available. We study this by quantifying the performance gap between training with data in a data-rich auxiliary language compared with training in the target language, exploring the benefits of translation systems, studying the limitations of model scaling when data is limited in the target languages, and proposing new methods for upsampling data from the auxiliary language. Our results show that stronger auxiliary datasets result in performance gains without modification to the model or training objective for close languages, and, in particular, that performance gains due to the development of more information-rich English pretraining datasets can extend to targeted language settings with limited data.