<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sogaard-2021-lockes">
    <titleInfo>
        <title>Locke’s Holiday: Belief Bias in Machine Reading</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Anders</namePart>
        <namePart type="family">Søgaard</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2021-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Online and Punta Cana, Dominican Republic</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <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 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 my kingdom, rather than 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.</abstract>
    <identifier type="citekey">sogaard-2021-lockes</identifier>
    <identifier type="doi">10.18653/v1/2021.emnlp-main.649</identifier>
    <location>
        <url>https://aclanthology.org/2021.emnlp-main.649</url>
    </location>
    <part>
        <date>2021-11</date>
        <extent unit="page">
            <start>8240</start>
            <end>8245</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Locke’s Holiday: Belief Bias in Machine Reading
%A Søgaard, Anders
%S Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
%D 2021
%8 November
%I Association for Computational Linguistics
%C Online and Punta Cana, Dominican Republic
%F sogaard-2021-lockes
%X 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 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 my kingdom, rather than 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.
%R 10.18653/v1/2021.emnlp-main.649
%U https://aclanthology.org/2021.emnlp-main.649
%U https://doi.org/10.18653/v1/2021.emnlp-main.649
%P 8240-8245

Markdown (Informal)

[Locke’s Holiday: Belief Bias in Machine Reading](https://aclanthology.org/2021.emnlp-main.649) (Søgaard, EMNLP 2021)

• Locke’s Holiday: Belief Bias in Machine Reading (Søgaard, EMNLP 2021)

ACL

• Anders Søgaard. 2021. Locke’s Holiday: Belief Bias in Machine Reading. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 8240–8245, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.