@inproceedings{muttenthaler-etal-2020-unsupervised,
title = "Unsupervised Evaluation for Question Answering with Transformers",
author = "Muttenthaler, Lukas and
Augenstein, Isabelle and
Bjerva, Johannes",
editor = "Alishahi, Afra and
Belinkov, Yonatan and
Chrupa{\l}a, Grzegorz and
Hupkes, Dieuwke and
Pinter, Yuval and
Sajjad, Hassan",
booktitle = "Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.blackboxnlp-1.8",
doi = "10.18653/v1/2020.blackboxnlp-1.8",
pages = "83--90",
abstract = "It is challenging to automatically evaluate the answer of a QA model at inference time. Although many models provide confidence scores, and simple heuristics can go a long way towards indicating answer correctness, such measures are heavily dataset-dependent and are unlikely to generalise. In this work, we begin by investigating the hidden representations of questions, answers, and contexts in transformer-based QA architectures. We observe a consistent pattern in the answer representations, which we show can be used to automatically evaluate whether or not a predicted answer span is correct. Our method does not require any labelled data and outperforms strong heuristic baselines, across 2 datasets and 7 domains. We are able to predict whether or not a model{'}s answer is correct with 91.37{\%} accuracy on SQuAD, and 80.7{\%} accuracy on SubjQA. We expect that this method will have broad applications, e.g., in semi-automatic development of QA datasets.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="muttenthaler-etal-2020-unsupervised">
<titleInfo>
<title>Unsupervised Evaluation for Question Answering with Transformers</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lukas</namePart>
<namePart type="family">Muttenthaler</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Isabelle</namePart>
<namePart type="family">Augenstein</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Johannes</namePart>
<namePart type="family">Bjerva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Afra</namePart>
<namePart type="family">Alishahi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yonatan</namePart>
<namePart type="family">Belinkov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Grzegorz</namePart>
<namePart type="family">Chrupała</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dieuwke</namePart>
<namePart type="family">Hupkes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuval</namePart>
<namePart type="family">Pinter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hassan</namePart>
<namePart type="family">Sajjad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>It is challenging to automatically evaluate the answer of a QA model at inference time. Although many models provide confidence scores, and simple heuristics can go a long way towards indicating answer correctness, such measures are heavily dataset-dependent and are unlikely to generalise. In this work, we begin by investigating the hidden representations of questions, answers, and contexts in transformer-based QA architectures. We observe a consistent pattern in the answer representations, which we show can be used to automatically evaluate whether or not a predicted answer span is correct. Our method does not require any labelled data and outperforms strong heuristic baselines, across 2 datasets and 7 domains. We are able to predict whether or not a model’s answer is correct with 91.37% accuracy on SQuAD, and 80.7% accuracy on SubjQA. We expect that this method will have broad applications, e.g., in semi-automatic development of QA datasets.</abstract>
<identifier type="citekey">muttenthaler-etal-2020-unsupervised</identifier>
<identifier type="doi">10.18653/v1/2020.blackboxnlp-1.8</identifier>
<location>
<url>https://aclanthology.org/2020.blackboxnlp-1.8</url>
</location>
<part>
<date>2020-11</date>
<extent unit="page">
<start>83</start>
<end>90</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Unsupervised Evaluation for Question Answering with Transformers
%A Muttenthaler, Lukas
%A Augenstein, Isabelle
%A Bjerva, Johannes
%Y Alishahi, Afra
%Y Belinkov, Yonatan
%Y Chrupała, Grzegorz
%Y Hupkes, Dieuwke
%Y Pinter, Yuval
%Y Sajjad, Hassan
%S Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP
%D 2020
%8 November
%I Association for Computational Linguistics
%C Online
%F muttenthaler-etal-2020-unsupervised
%X It is challenging to automatically evaluate the answer of a QA model at inference time. Although many models provide confidence scores, and simple heuristics can go a long way towards indicating answer correctness, such measures are heavily dataset-dependent and are unlikely to generalise. In this work, we begin by investigating the hidden representations of questions, answers, and contexts in transformer-based QA architectures. We observe a consistent pattern in the answer representations, which we show can be used to automatically evaluate whether or not a predicted answer span is correct. Our method does not require any labelled data and outperforms strong heuristic baselines, across 2 datasets and 7 domains. We are able to predict whether or not a model’s answer is correct with 91.37% accuracy on SQuAD, and 80.7% accuracy on SubjQA. We expect that this method will have broad applications, e.g., in semi-automatic development of QA datasets.
%R 10.18653/v1/2020.blackboxnlp-1.8
%U https://aclanthology.org/2020.blackboxnlp-1.8
%U https://doi.org/10.18653/v1/2020.blackboxnlp-1.8
%P 83-90
Markdown (Informal)
[Unsupervised Evaluation for Question Answering with Transformers](https://aclanthology.org/2020.blackboxnlp-1.8) (Muttenthaler et al., BlackboxNLP 2020)
ACL