CORAL: Adaptive Retrieval Loop for Culturally-Aligned Multilingual RAG

TL;DR

CORAL introduces an adaptive, iterative retrieval method for multilingual RAG that refines both the corpus and query to improve cultural relevance and answer accuracy.

Contribution

It proposes a novel context-aware retrieval loop that enhances cultural alignment in multilingual retrieval-augmented generation systems.

Findings

Up to 3.58% accuracy improvement on low-resource languages.

CORAL effectively refines retrieval and query rewriting for cultural relevance.

Demonstrates superiority over strong baselines on cultural QA benchmarks.

Abstract

Multilingual retrieval-augmented generation (mRAG) is often implemented within a fixed retrieval space, typically via query or document translation or multilingual embedding vector representations. However, this approach may be inadequate for culturally grounded queries, in which retrieval-condition misalignment may occur. Even strong retrievers and generators may struggle to produce culturally relevant answers when sourcing evidence from inappropriate linguistic or regional contexts. To this end, we introduce CORAL (COntext-aware Retrieval with Agentic Loop, an adaptive retrieval methodology for mRAG that enables iterative refinement of both the retrieval space (corpora) and the retrieval probe (query) based on the quality of the evidence. The overall process includes: (1) selecting corpora, (2) retrieving documents, (3) critiquing evidence for relevance and cultural alignment, and (4) checking sufficiency. If the retrieved documents are insufficient to answer the query correctly, the system (5) reselects corpora and rewrites the query. Across two cultural QA benchmarks, CORAL achieves up to a 3.58%p accuracy improvement on low-resource languages relative to the strongest baselines.