Impact of Artificially Code-Switched Training Data Volume on Zero-Shot Retrieval for Low-Resource Languages in MIRACL
Domaine:
natural language processingType de record:
paper
Créateur:
Ass
Éditeur:
Hôte:
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: What is the impact of scaling the volume of artificially code-switched training data on the zero-shot retrieval performance for low-resource languages in the MIRACL dataset?
Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.5/10. This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 7.5/10.
Similaires
Artificially Code-Switched Training Data Volume and Zero-Shot MIRACL Retrieval Accuracy Across Language Resource LevelsArtificially Code-Switched Training for Zero-Shot Cross-Lingual Retrieval in Low-Resource LanguagesScaling Artificially Code-Switched Training Data for Zero-Shot Multilingual Dense Retrieval in Low-Resource LanguagesScaling Bilingual Lexicons for Zero-Shot Cross-Lingual Retrieval on Artificially Code-Switched Low-Resource Data in MIRACLScaling Artificially Code-Switched Data for Zero-Shot Retrieval in Low-Resource Afro-Asiatic LanguagesCross-Lingual Embeddings for Zero-Shot Retrieval on Artificially Code-Switched Low-Resource Data