Zero-Shot Performance in XTREME-M with Intermediate-Task English Training vs. Direct Multimodal Fine-Tuning

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: How does English intermediate-task training impact zero-shot performance on the XTREME-M image-text retrieval benchmark compared to direct fine-tuning on multimodal transformers like M-UNITER?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.7/10.