Comparison of Inference Latency in Multimodal vs. Text-Only Intermediate Tasks for Zero-Shot Cross-Lingual Transfer on 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: How does the inference latency compare between multimodal intermediate tasks and text-only intermediate tasks in zero-shot cross-lingual transfer on the XTREME benchmark when using models like BERT or XLM-R, and what trade-offs exist between latency and accuracy?

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