Multilingual vs. English-Only Intermediate Tasks in Zero-Shot Cross-Lingual Transfer with Adversarial Perturbations
Year: 2026
Type: report
Abstract: 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 performance of multilingual intermediate tasks compare to English-only intermediate tasks in zero-shot cross-lingual transfer when evaluated on XTREME-R with adversarial perturbations, measured by accuracy drops and robustness metrics? Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.4/10.
Author: Assignee Research
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204708986
Cites:
Cited by:
Related to: