@inproceedings{chen-etal-2022-outlier,
    title = "Outlier-Aware Training for Improving Group Accuracy Disparities",
    author = "Chen, Li-Kuang  and
      Kruengkrai, Canasai  and
      Yamagishi, Junichi",
    editor = "Hanqi, Yan  and
      Zonghan, Yang  and
      Ruder, Sebastian  and
      Xiaojun, Wan",
    booktitle = "Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing: Student Research Workshop",
    month = nov,
    year = "2022",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2022.aacl-srw.8/",
    doi = "10.18653/v1/2022.aacl-srw.8",
    pages = "54--60",
    abstract = "Methods addressing spurious correlations such as Just Train Twice (JTT, Liu et al. 2021) involve reweighting a subset of the training set to maximize the worst-group accuracy. However, the reweighted set of examples may potentially contain unlearnable examples that hamper the model{'}s learning. We propose mitigating this by detecting outliers to the training set and removing them before reweighting. Our experiments show that our method achieves competitive or better accuracy compared with JTT and can detect and remove annotation errors in the subset being reweighted in JTT."
}