https://github.com/chenzhongwu20/RuleRAG_ICL_FT

RuleRAG: Rule-guided Retrieval-Augmented Generation with Language Models for Question Answering
https://github.com/chenzhongwu20/RuleRAG_ICL_FT

Last synced: 12 days ago
JSON representation

RuleRAG: Rule-guided Retrieval-Augmented Generation with Language Models for Question Answering

• Host: GitHub
• URL: https://github.com/chenzhongwu20/RuleRAG_ICL_FT
• Owner: chenzhongwu20
• Created: 2024-10-02T04:09:24.000Z (4 months ago)
• Default Branch: main
• Last Pushed: 2024-11-15T04:54:14.000Z (3 months ago)
• Last Synced: 2024-11-15T05:27:30.648Z (3 months ago)
• Language: Python
• Homepage:
• Size: 2.13 MB
• Stars: 10
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesomerag_paper - https://github.com/chenzhongwu20/RuleRAG_ICL_FT
• awesomerag_paper - https://github.com/chenzhongwu20/RuleRAG_ICL_FT

README

        
**RuleRAG: Rule-guided Retrieval-Augmented Generation with Language Models for Question Answering**

====

****

This is the official implementation repository of the paper, RuleRAG: Rule-guided retrieval-augmented generation with language models for question answering.

 

****

![image](framework.png)

The framework of our proposed RuleRAG, including RuleRAG-ICL and RuleRAG-FT. RuleRAG-ICL relies on in-context learning (ICL) with the guidance of rules. RuleRAG-FT involves fine-tuning retrievers and generators ahead.

 

****

Our constructed five RuleQA benchmarks are stored on [Google Drive](https://drive.google.com/drive/folders/13tbJS-Eq3Cswck3JRPU0LJIZ1Vrz3Bga?usp=sharing).

![image](data.png)

The statistics of the five RuleQA benchmarks. 

 

****

Trian the retrievers of RuleRAG: We use the public implementation of the retrievers. The training data is stored on [Google Drive](https://drive.google.com/drive/folders/13tbJS-Eq3Cswck3JRPU0LJIZ1Vrz3Bga?usp=sharing)

Train the generators of RuleRAG: main_train.py

Inference of RuleRAG: RuleRAG_inference.py