@inproceedings{mu-lim-2022-universal,
title = "Universal Evasion Attacks on Summarization Scoring",
author = "Mu, Wenchuan and
Lim, Kwan Hui",
editor = "Bastings, Jasmijn and
Belinkov, Yonatan and
Elazar, Yanai and
Hupkes, Dieuwke and
Saphra, Naomi and
Wiegreffe, Sarah",
booktitle = "Proceedings of the Fifth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates (Hybrid)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.blackboxnlp-1.9",
doi = "10.18653/v1/2022.blackboxnlp-1.9",
pages = "104--118",
abstract = "The automatic scoring of summaries is important as it guides the development of summarizers. Scoring is also complex, as it involves multiple aspects such as the fluency, grammar, and even textual entailment with the source text. However, summary scoring has not been considered as a machine learning task to study its accuracy and robustness. In this study, we place automatic scoring in the context of regression machine learning tasks and perform evasion attacks to explore its robustness. Attack systems predict a non-summary string from each input, and these non-summary strings achieve competitive scores with good summarizers on the most popular metrics: ROUGE, METEOR, and BERTScore. Attack systems also {``}outperform{''} state-of-the-art summarization methods on ROUGE-1 and ROUGE-L, and score the second-highest on METEOR. Furthermore, a BERTScore backdoor is observed: a simple trigger can score higher than any automatic summarization method. The evasion attacks in this work indicate the low robustness of current scoring systems at the system level. We hope that our highlighting of these proposed attack will facilitate the development of summary scores.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mu-lim-2022-universal">
<titleInfo>
<title>Universal Evasion Attacks on Summarization Scoring</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wenchuan</namePart>
<namePart type="family">Mu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kwan</namePart>
<namePart type="given">Hui</namePart>
<namePart type="family">Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fifth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jasmijn</namePart>
<namePart type="family">Bastings</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">Yanai</namePart>
<namePart type="family">Elazar</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">Naomi</namePart>
<namePart type="family">Saphra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sarah</namePart>
<namePart type="family">Wiegreffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, United Arab Emirates (Hybrid)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The automatic scoring of summaries is important as it guides the development of summarizers. Scoring is also complex, as it involves multiple aspects such as the fluency, grammar, and even textual entailment with the source text. However, summary scoring has not been considered as a machine learning task to study its accuracy and robustness. In this study, we place automatic scoring in the context of regression machine learning tasks and perform evasion attacks to explore its robustness. Attack systems predict a non-summary string from each input, and these non-summary strings achieve competitive scores with good summarizers on the most popular metrics: ROUGE, METEOR, and BERTScore. Attack systems also “outperform” state-of-the-art summarization methods on ROUGE-1 and ROUGE-L, and score the second-highest on METEOR. Furthermore, a BERTScore backdoor is observed: a simple trigger can score higher than any automatic summarization method. The evasion attacks in this work indicate the low robustness of current scoring systems at the system level. We hope that our highlighting of these proposed attack will facilitate the development of summary scores.</abstract>
<identifier type="citekey">mu-lim-2022-universal</identifier>
<identifier type="doi">10.18653/v1/2022.blackboxnlp-1.9</identifier>
<location>
<url>https://aclanthology.org/2022.blackboxnlp-1.9</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>104</start>
<end>118</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Universal Evasion Attacks on Summarization Scoring
%A Mu, Wenchuan
%A Lim, Kwan Hui
%Y Bastings, Jasmijn
%Y Belinkov, Yonatan
%Y Elazar, Yanai
%Y Hupkes, Dieuwke
%Y Saphra, Naomi
%Y Wiegreffe, Sarah
%S Proceedings of the Fifth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates (Hybrid)
%F mu-lim-2022-universal
%X The automatic scoring of summaries is important as it guides the development of summarizers. Scoring is also complex, as it involves multiple aspects such as the fluency, grammar, and even textual entailment with the source text. However, summary scoring has not been considered as a machine learning task to study its accuracy and robustness. In this study, we place automatic scoring in the context of regression machine learning tasks and perform evasion attacks to explore its robustness. Attack systems predict a non-summary string from each input, and these non-summary strings achieve competitive scores with good summarizers on the most popular metrics: ROUGE, METEOR, and BERTScore. Attack systems also “outperform” state-of-the-art summarization methods on ROUGE-1 and ROUGE-L, and score the second-highest on METEOR. Furthermore, a BERTScore backdoor is observed: a simple trigger can score higher than any automatic summarization method. The evasion attacks in this work indicate the low robustness of current scoring systems at the system level. We hope that our highlighting of these proposed attack will facilitate the development of summary scores.
%R 10.18653/v1/2022.blackboxnlp-1.9
%U https://aclanthology.org/2022.blackboxnlp-1.9
%U https://doi.org/10.18653/v1/2022.blackboxnlp-1.9
%P 104-118
Markdown (Informal)
[Universal Evasion Attacks on Summarization Scoring](https://aclanthology.org/2022.blackboxnlp-1.9) (Mu & Lim, BlackboxNLP 2022)
ACL
- Wenchuan Mu and Kwan Hui Lim. 2022. Universal Evasion Attacks on Summarization Scoring. In Proceedings of the Fifth BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP, pages 104–118, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics.