Improving Zero-Shot Cross-Lingual Transfer for Programming Tasks via Intermediate-Task Training

Pre-trained multilingual language encoders, such as multilingual BERT and XLM-R, show great potential for zero-shot cross-lingual transfer. However, these multilingual encoders do not precisely align words and phrases across languages. Especially, learning alignments in the multilingual embedding space usually requires sentence-level or word-level parallel corpora, which are expensive to be obtained for low-resource languages. An alternative is to make the multilingual encoders more robust; when fine-tuning the encoder using downstream task, we train the encoder to tolerate noise in the contex

Research goal: Can intermediate-task training on code-related tasks (e.g., code generation or code-to-text) improve zero-shot cross-lingual transfer for programming-related tasks in XTREME-R benchmarks?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.7/10.