Multilingual Information Retrieval with a Monolingual Knowledge Base

2025-06-03

AI Summary

This paper addresses the challenge of effectively leveraging monolingual knowledge bases for low-resource languages in cross-lingual knowledge transfer. We propose a language-agnostic weighted sampling contrastive learning fine-tuning method: multilingual sentences are jointly mapped into the embedding space of a pivot language (e.g., English) without requiring translation or bilingual alignment, enabling cross-lingual semantic alignment. The method is inherently compatible with multilingual mixing and code-switching scenarios and relies solely on monolingual knowledge bases. Evaluated on cross-lingual information retrieval, it achieves a 31.03% improvement in MRR and a 33.98% gain in Recall@3 over standard sampling strategies. These results demonstrate substantial gains in retrieval effectiveness for low-resource languages. Our approach establishes a new paradigm—efficient, lightweight, and scalable—for cross-lingual knowledge acquisition, circumventing reliance on parallel data or resource-intensive translation pipelines.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Multilingual information retrieval has emerged as powerful tools for expanding knowledge sharing across languages. On the other hand, resources on high quality knowledge base are often scarce and in limited languages, therefore an effective embedding model to transform sentences from different languages into a feature vector space same as the knowledge base language becomes the key ingredient for cross language knowledge sharing, especially to transfer knowledge available in high-resource languages to low-resource ones. In this paper we propose a novel strategy to fine-tune multilingual embedding models with weighted sampling for contrastive learning, enabling multilingual information retrieval with a monolingual knowledge base. We demonstrate that the weighted sampling strategy produces performance gains compared to standard ones by up to 31.03% in MRR and up to 33.98% in Recall@3. Additionally, our proposed methodology is language agnostic and applicable for both multilingual and code switching use cases.

Problem

• Enabling cross-language knowledge sharing with monolingual knowledge base
• Improving multilingual embedding models for information retrieval
• Transferring knowledge from high-resource to low-resource languages

Innovation

• Fine-tune multilingual embedding models
• Weighted sampling for contrastive learning
• Enable retrieval with monolingual knowledge base