Enhancing Multilingual RAG Systems with Debiased Language Preference-Guided Query Fusion

TL;DR

This paper identifies biases in multilingual RAG systems that favor English due to structural evaluation biases and proposes DeLP and DELTA methods to mitigate these biases, improving cross-lingual retrieval and generation.

Contribution

The paper introduces DeLP, a debiased metric for assessing language preference, and DELTA, a framework leveraging monolingual alignment to enhance multilingual RAG performance.

Findings

DeLP reveals that English preference is due to evidence distribution, not inherent bias.

Retrievers favor monolingual query-document alignment.

DELTA outperforms baseline methods across multiple languages.

Abstract

Multilingual Retrieval-Augmented Generation (mRAG) systems often exhibit a perceived preference for high-resource languages, particularly English, resulting in the widespread adoption of English pivoting. While prior studies attribute this advantage to the superior English-centric capabilities of Large Language Models (LLMs), we find that such measurements are significantly distorted by structural priors inherent in evaluation benchmarks. Specifically, we identify exposure bias and a gold availability prior-both driven by the disproportionate concentration of resources in English-as well as cultural priors rooted in topic locality, as factors that hinder accurate assessment of genuine language preference. To address these biases, we propose DeLP (Debiased Language Preference), a calibrated metric designed to explicitly factor out these structural confounds. Our analysis using DeLP reveals that the previously reported English preference is largely a byproduct of evidence distribution rather than an inherent model bias. Instead, we find that retrievers fundamentally favor monolingual alignment between the query and the document language. Building on this insight, we introduce DELTA (DEbiased Language preference-guided Text Augmentation), a lightweight and efficient mRAG framework that strategically leverages monolingual alignment to optimize cross-lingual retrieval and generation. Experimental results demonstrate that DELTA consistently outperforms English pivoting and mRAG baselines across diverse languages.