XLM-R-large Performance in Zero-Shot Cross-Lingual NER: Task Complexity Effects

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 choice of intermediate task complexity (e.g., simple classification vs. complex structured prediction tasks) impact the zero-shot cross-lingual performance of XLM-R-large on XTREME-R's NER task, measured by F1 score comparisons across varying task difficulties?

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