Multimodal Intermediate-Task Training for Zero-Shot Cross-Lingual Transfer in XTREME-R Vision-Language Subsets

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: To what extent does intermediate-task training on multimodal reasoning datasets enhance zero-shot cross-lingual transfer on vision-language subsets of XTREME-R relative to text-only English intermediate training?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.5/10.