@inproceedings{sogaard-2021-lockes,
    title = "Locke{'}s Holiday: Belief Bias in Machine Reading",
    author = "S{\o}gaard, Anders",
    editor = "Moens, Marie-Francine  and
      Huang, Xuanjing  and
      Specia, Lucia  and
      Yih, Scott Wen-tau",
    booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2021",
    address = "Online and Punta Cana, Dominican Republic",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2021.emnlp-main.649/",
    doi = "10.18653/v1/2021.emnlp-main.649",
    pages = "8240--8245",
    abstract = "I highlight a simple failure mode of state-of-the-art machine reading systems: when contexts do not align with commonly shared beliefs. For example, machine reading systems fail to answer \textit{What did Elizabeth want?} correctly in the context of `My kingdom for a cough drop, cried Queen Elizabeth.' Biased by co-occurrence statistics in the training data of pretrained language models, systems predict \textit{my kingdom}, rather than \textit{a cough drop}. I argue such biases are analogous to human belief biases and present a carefully designed challenge dataset for English machine reading, called Auto-Locke, to quantify such effects. Evaluations of machine reading systems on Auto-Locke show the pervasiveness of belief bias in machine reading."
}