This the the repository for the ACL-2021 long paper --- Counterfactual Inference for Text Classification Debiasing.
By leveraging the causal intervention, we propose a model-agnostic text classification debiasing framework – CORSAIR, which can effectively avoid employing data manipulations or designing balancing mechanisms.
- code/ contains the source codes.
- data/ contains eleven datasets used for evaluating.
- Python (≥3.0)
- PyTorch (≥1.0)
- All hyperparameters are in _public.py.
If you find this study helpful or related, please kindly consider citing as:
@inproceedings{Corsair,
title = {Counterfactual Inference for Text Classification Debiasing},
author = {Chen Qian and Fuli Feng and Lijie Wen and Chunping Ma and Pengjun Xie},
booktitle = {Proceedings of the Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021)},
year = {2021},
pages = {5434–5445}
}