@inproceedings{zylberajch-etal-2021-hildif,
    title = "{HILDIF}: {I}nteractive Debugging of {NLI} Models Using Influence Functions",
    author = "Zylberajch, Hugo  and
      Lertvittayakumjorn, Piyawat  and
      Toni, Francesca",
    editor = {Brantley, Kiant{\'e}  and
      Dan, Soham  and
      Gurevych, Iryna  and
      Lee, Ji-Ung  and
      Radlinski, Filip  and
      Sch{\"u}tze, Hinrich  and
      Simpson, Edwin  and
      Yu, Lili},
    booktitle = "Proceedings of the First Workshop on Interactive Learning for Natural Language Processing",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.internlp-1.1",
    doi = "10.18653/v1/2021.internlp-1.1",
    pages = "1--6",
    abstract = "Biases and artifacts in training data can cause unwelcome behavior in text classifiers (such as shallow pattern matching), leading to lack of generalizability. One solution to this problem is to include users in the loop and leverage their feedback to improve models. We propose a novel explanatory debugging pipeline called HILDIF, enabling humans to improve deep text classifiers using influence functions as an explanation method. We experiment on the Natural Language Inference (NLI) task, showing that HILDIF can effectively alleviate artifact problems in fine-tuned BERT models and result in increased model generalizability.",
}