SOAS English-Uzbek RAG Evaluation (Retrieval-Only)
Abstract
Bilingual English-Uzbek retrieval-evaluation benchmark for culturally grounded retrieval-augmented generation systems. This retrieval-only release contains 400 public evaluation rows, balanced across English and Uzbek and across four domains: governance, history, institutions, and culture. Each row includes a question, language, domain, source document identifiers, source title, answerability, cultural-specificity label, difficulty label, and quality flag. The release excludes answer text, reference answers, source excerpts, and full source documents. Headline result from the associated experiments: targeted Uzbek source curation improved Uzbek retrieval recall from 39% to 98% (Cohen's d = 2.91), suggesting that corpus coverage can matter more than model choice for low-resource and culturally grounded RAG evaluation.
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.