Task Diversity in Intermediate Training and Cross-Lingual Zero-Shot Performance

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: What is the impact of task diversity (e.g., question answering vs. natural language inference) in intermediate-task training on cross-lingual zero-shot performance, as measured by accuracy on XTREME tasks, and how does this relate to the semantic similarity between intermediate and target tasks?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.2/10.