Scaling of Source Languages and Cross-Lingual Transfer in Multi-Level Contrastive Models

Multi-lingual language models (LM), such as mBERT, XLM-R, mT5, mBART, have been remarkably successful in enabling natural language tasks in low-resource languages through cross-lingual transfer from high-resource ones. In this work, we try to better understand how such models, specifically mT5, transfer *any* linguistic and semantic knowledge across languages, even though no explicit cross-lingual signals are provided during pre-training. Rather, only unannotated texts from each language are presented to the model separately and independently of one another, and the model appears to implicitly

Research goal: What is the scaling relationship between the number of source languages in the training set and the cross-lingual transfer performance of multi-level contrastive models on low-resource slots in X-TREME-SLU?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.5/10.