Multimodal Intermediate-Task Training for Zero-Shot Cross-Lingual Transfer in XTREME

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 multimodal intermediate-task training (e.g., image-text retrieval) improve zero-shot cross-lingual transfer performance on XTREME compared to purely textual intermediate tasks, and how does this vary across classification and retrieval tasks?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.8/10.