This is the official repository of IterCQR: Iterative Conversational Query Reformulation with Retrieval Guidance.
This paper has been accepted for NAACL 2024.
pip install -r requirements.txthttps://github.com/fengranMark/ConvGQR/tree/main
Follow the instructions provided in that repository for data preprocessing.
For IterCQR model initialization, we use a dataset generated by gpt-3.5-turbo.You can download the dataset from the following link:
https://drive.google.com/drive/folders/1ABCdVlZWYgBe-JNmbPnp5CZ-W0-fucCV?usp=drive_link
Refer to the provided bash scripts to train the model for TopiOCQA and QReCC. We release our trained model at the following link:https://drive.google.com/drive/folders/1ABCdVlZWYgBe-JNmbPnp5CZ-W0-fucCV?usp=drive_link
If you use IterCQR in your research, please cite our work:@inproceedings{jang2024itercqr,
title={IterCQR: Iterative Conversational Query Reformulation with Retrieval Guidance},
author={Jang, Yunah and Lee, Kang-il and Bae, Hyunkyung and Lee, Hwanhee and Jung, Kyomin},
booktitle={Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)},
pages={8114--8131},
year={2024}
}
'''