Domain-Specific Intermediate-Task Training for Robust Zero-Shot Cross-Lingual Transfer on Adversarial Inputs

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 domain-specific intermediate-task training (e.g., legal, medical) in English lead to higher robustness in zero-shot cross-lingual transfer on XTREME for adversarially perturbed inputs, measured by accuracy drops on perturbed vs. unperturbed test sets?

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