PersianRAG: A Retrieval-Augmented Generation System for Persian Language

TL;DR

PersianRAG introduces a retrieval-augmented generation system tailored for Persian, overcoming language-specific challenges to improve question answering performance in low-resource settings.

Contribution

The paper presents novel solutions for implementing RAG models in Persian, addressing unique preprocessing, retrieval, and evaluation challenges.

Findings

Enhanced question answering accuracy in Persian.

Effective handling of Persian language challenges in RAG models.

Demonstrated improvements on Persian benchmark datasets.

Abstract

Retrieval augmented generation (RAG) models, which integrate large-scale pre-trained generative models with external retrieval mechanisms, have shown significant success in various natural language processing (NLP) tasks. However, applying RAG models in Persian language as a low-resource language, poses distinct challenges. These challenges primarily involve the preprocessing, embedding, retrieval, prompt construction, language modeling, and response evaluation of the system. In this paper, we address the challenges towards implementing a real-world RAG system for Persian language called PersianRAG. We propose novel solutions to overcome these obstacles and evaluate our approach using several Persian benchmark datasets. Our experimental results demonstrate the capability of the PersianRAG framework to enhance question answering task in Persian.