Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

Robustness of mT5 and XLM-R versus Monolingual Code-Switched Models in Zero-Shot Low-Resource Retrieval

Domain:

natural language processing

Record type:

paper
Creator:
Ass
Publisher:
Zenodo
Host:avatar
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 robustness of mT5 and XLM-R in zero-shot cross-lingual retrieval compare to monolingual models trained on code-switched data when evaluated on low-resource language pairs using nDCG@10? Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.3/10. This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 8.3/10.

Visit

doi.orgzenodo.org

Tasks

code switchinginformation retrieval

Tags

robustnessmT5XLM-Rzero-shotcross-lingualretrievalmonolingualmodels

Licenses

Creative Commons Attribution 4.0 Internationalhttps://creativecommons.org/licenses/by/4.0/legalcode

Similar

Robustness of Zero-Shot Cross-Lingual Retrieval Models via Code-Switched Pre-Training in Low-Resource LanguagesZero-shot Cross-lingual Retrieval Accuracy of XLM-R and mBART on Code-switched Low-resource Language PairsmT5 Code-Switching Training for Low-Resource Zero-Shot RetrievalMultilingual BERT versus XLM-R for Zero-Shot Cross-Lingual Retrieval on Low-Resource Pairs with Artificial Code-SwitchingZero-Shot Cross-Lingual Retrieval Robustness Under Varying Low-Resource Language Proportions in Code-Switched Training DataComparative Analysis of Hybrid Batch Training Against XLM-R and mT5 for Zero-Shot Cross-Lingual Retrieval in Low-Resource

Robustness of Zero-Shot Cross-Lingual Retrieval Models via Code-Switched Pre-Training in Low-Resource Languages

Transferring information retrieval (IR) models from a high-resource language (typically English) to

Zero-shot Cross-lingual Retrieval Accuracy of XLM-R and mBART on Code-switched Low-resource Language Pairs

Transferring information retrieval (IR) models from a high-resource language (typically English) to

mT5 Code-Switching Training for Low-Resource Zero-Shot Retrieval

Transferring information retrieval (IR) models from a high-resource language (typically English) to

Multilingual BERT versus XLM-R for Zero-Shot Cross-Lingual Retrieval on Low-Resource Pairs with Artificial Code-Switching

Transferring information retrieval (IR) models from a high-resource language (typically English) to

Zero-Shot Cross-Lingual Retrieval Robustness Under Varying Low-Resource Language Proportions in Code-Switched Training Data

Benefiting from transformer-based pre-trained language models, neural ranking models have made signi

Comparative Analysis of Hybrid Batch Training Against XLM-R and mT5 for Zero-Shot Cross-Lingual Retrieval in Low-Resource

Information retrieval across different languages is an increasingly important challenge in natural l