Bilingual Lexicon Quality from Static and Contextual Embeddings for Zero-Shot Cross-Lingual Retrieval on MIRACL

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 does the quality of bilingual lexicons derived from different sources (e.g., static word embeddings vs. contextualized embeddings) impact the effectiveness of artificially code-switched data in improving zero-shot cross-lingual retrieval on MIRACL and other multilingual benchmarks like BEIR or NQ, measured by nDCG@10?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.1/10.