Scaling Multilingual versus Monolingual Intermediate-Task Training for Downstream Performance on XTREME-UDA

Intermediate-task training---fine-tuning a pretrained model on an intermediate task before fine-tuning again on the target task---often improves model performance substantially on language understanding tasks in monolingual English settings. We investigate whether English intermediate-task training is still helpful on non-English target tasks. Using nine intermediate language-understanding tasks, we evaluate intermediate-task transfer in a zero-shot cross-lingual setting on the XTREME benchmark. We see large improvements from intermediate training on the BUCC and Tatoeba sentence retrieval tas

Research goal: Does scaling the size of the pretrained model (e.g., from XLM-R_base to XLM-R_large) while using multilingual intermediate-task training improve downstream performance on XTREME-UDA, and how does this compare to monolingual English intermediate-task scaling?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.3/10.