This repository is the Pytorch implementation of our paper " Introspective Distillation for Robust Question Answering" in NeurIPS 2021.
IntroD is proposed to achieve both high in-distribution (ID) and out-of-distribution (OOD) performances for question answering tasks like VQA and extractive QA. The key technical contribution is to blend the inductive bias of OOD and ID by introspecting whether a training sample fits in the factual ID world or the counterfactual OOD one.
If you find this paper and codes help your research, please kindly consider citing our papers in your publications.
@inproceedings{niu2021introspective,
title={Introspective Distillation for Robust Question Answering},
author={Niu, Yulei and Zhang, Hanwang},
booktitle={Thirty-Fifth Conference on Neural Information Processing Systems},
year={2021}
}
@inproceedings{niu2020counterfactual,
title={Counterfactual VQA: A Cause-Effect Look at Language Bias},
author={Niu, Yulei and Tang, Kaihua and Zhang, Hanwang and Lu, Zhiwu and Hua, Xian-Sheng and Wen, Ji-Rong},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year={2021}
}