Robustness of Zero-Shot Cross-Lingual Retrieval Models Trained on Code-Switched Data Across Unseen Language Pairs and Domains

Transferring information retrieval (IR) models from a high-resource language (typically English) to other languages in a zero-shot fashion has become a widely adopted approach. In this work, we show that the effectiveness of zero-shot rankers diminishes when queries and documents are present in different languages. Motivated by this, we propose to train ranking models on artificially code-switched data instead, which we generate by utilizing bilingual lexicons. To this end, we experiment with lexicons induced from (1) cross-lingual word embeddings and (2) parallel Wikipedia page titles. We use

Research goal: How robust are zero-shot cross-lingual retrieval models trained on code-switched data when evaluated on unseen language pairs or domains (e.g., technical vs. conversational queries), as measured by accuracy drops in Recall@50 compared to fine-tuned baselines?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.