FairRAG: A Privacy-Preserving Framework for Fair Financial Decision-Making

R. Nagpal, U. Usua, R. Palacios, A. Gupta

Applied Sciences Vol. 15, nº. 15, pp. 8282

Summary:

Customer churn prediction has become crucial for businesses, yet it poses significant challenges regarding privacy preservation and prediction accuracy. In this paper, we address two fundamental questions: (1) How can customer churn be effectively predicted while ensuring robust privacy protection of sensitive data? (2) How can large language models enhance churn prediction accuracy while maintaining data privacy? To address these questions, we propose FairRAG, a robust architecture that combines differential privacy, retrieval-augmented generation, and LLMs. Our approach leverages OPT-125M as the core language model along with a sentence transformer for semantic similarity matching while incorporating differential privacy mechanisms to generate synthetic training data. We evaluate FairRAG on two diverse datasets: Bank Churn and Telco Churn. The results demonstrate significant improvements over both traditional machine learning approaches and standalone LLMs, achieving accuracy improvements of up to 11% on the Bank Churn dataset and 12% on the Telco Churn dataset. These improvements were maintained when using differentially private synthetic data, thus indicating robust privacy and accuracy trade-offs.

Spanish layman's summary:

FairRAG es una arquitectura robusta que combina privacidad diferencial, generación aumentada por recuperación y modelos de lenguaje. Los resultados muestran mejoras significativas frente a enfoques de aprendizaje automático y LLMs independientes, incluso con datos sintéticos privados.

English layman's summary:

FairRAG is a robust architecture that combines differential privacy, retrieval-augmented generation, and LLMs. The results show significant improvement over machine learning approaches and standalone LLMs even with differentially private synthetic data.

Keywords: algorithmic fairness; privacy-preserving machine learning; differential privacy; retrieval-augmented generation

JCR-JIF Impact Factor and WoS quartile: 2,900 - Q2 (2025)

DOI reference: https://doi.org/10.3390/app15158282

Published on paper: August 2025.

Published on-line: July 2025.

Citation:
R. Nagpal, U. Usua, R. Palacios, A. Gupta, "FairRAG: A Privacy-Preserving Framework for Fair Financial Decision-Making", Applied Sciences, Vol. 15, nº. 15, pp. 8282, August 2025. [Online: July 2025] doi: 10.3390/app15158282