Skip to content

This is the code repo for our paper "Revealing the Treasures of Knowledge via Active Learning".

License

Notifications You must be signed in to change notification settings

OpenMatch/ActiveRAG

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ActiveRAG: Revealing the Treasures of Knowledge via Active Learning

Source code for our paper :
ActiveRAG: Revealing the Treasures of Knowledge via Active Learning

If you find this work useful, please cite our paper and give us a shining star 🌟

Overview

ActiveRAG is an innovative RAG framework that shifts from passive knowledge acquisition to an active learning mechanism. This approach utilizes the Knowledge Construction mechanism to develop a deeper understanding of external knowledge by associating it with previously acquired or memorized knowledge. Subsequently, it designs the Cognitive Nexus mechanism to incorporate the outcomes from both chains of thought and knowledge construction, thereby calibrating the intrinsic cognition of LLMs.

ActiveRAG

Quick Start

Install from git

git clone https://github.com/OpenMatch/ActiveRAG
pip install -r requirements.txt

Reproduction

We provide our request logs, so the results in the paper can be quickly reproduced:

python -m logs.eval --dataset nq --topk 5

Parameters:

  • dataset: dataset name.
  • topk: using top-k of retrieved passages to augment.

Re-request

We also provide the full request code, you can re-request for further exploration.

First, set your own api-key in agent file:

openai.api_key = 'sk-<your-api-key>'

Then, run the following script:

python -m scripts.run --dataset nq --topk 5

Analyzing log files:

python -m scripts.build --dataset nq --topk 5

Evaluate:

python -m scripts.evaluate --dataset nq --topk 5

Citation

@article{xu2024activerag,
  title={ActiveRAG: Revealing the Treasures of Knowledge via Active Learning},
  author={Xu, Zhipeng and Liu, Zhenghao and Liu, Yibin and Xiong, Chenyan and Yan, Yukun and Wang, Shuo and Yu, Shi and Liu, Zhiyuan and Yu, Ge},
  journal={arXiv preprint arXiv:2402.13547},
  year={2024}
}

Contact Us

If you have questions, suggestions, and bug reports, please send a email to us, we will try our best to help you.

About

This is the code repo for our paper "Revealing the Treasures of Knowledge via Active Learning".

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages