Domain-Specific Intermediate Task Integration for Zero-Shot Cross-Lingual Transfer in XTREME-R
Description
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 effect of combining domain-specific intermediate tasks (e.g., legal, biomedical) on zero-shot cross-lingual transfer in XTREME-R compared to general language understanding tasks?
Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.8/10.
Notes
Files
paper.pdf
Files
(78.7 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:c412d88756f4a1c469311d9d1cfe73f9
|
78.7 kB | Preview Download |