@inproceedings{raunak-etal-2022-salted,
title = "{SALTED}: A Framework for {SA}lient Long-tail Translation Error Detection",
author = "Raunak, Vikas and
Post, Matt and
Menezes, Arul",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-emnlp.379",
doi = "10.18653/v1/2022.findings-emnlp.379",
pages = "5163--5179",
abstract = "Traditional machine translation (MT) metrics provide an average measure of translation quality that is insensitive to the long tail of behavioral problems. Examples include translation of numbers, physical units, dropped content and hallucinations. These errors, which occur rarely and unpredictably in Neural Machine Translation (NMT), greatly undermine the reliability of state-of-the-art MT systems. Consequently, it is important to have visibility into these problems during model development. Towards this end, we introduce SALTED, a specifications-based framework for behavioral testing of NMT models. At the core of our approach is the use of high-precision detectors that flag errors (or alternatively, verify output correctness) between a source sentence and a system output. These detectors provide fine-grained measurements of long-tail errors, providing a trustworthy view of problems that were previously invisible. We demonstrate that such detectors could be used not just to identify salient long-tail errors in MT systems, but also for higher-recall filtering of the training data, fixing targeted errors with model fine-tuning in NMT and generating novel data for metamorphic testing to elicit further bugs in models.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="raunak-etal-2022-salted">
<titleInfo>
<title>SALTED: A Framework for SAlient Long-tail Translation Error Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vikas</namePart>
<namePart type="family">Raunak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Matt</namePart>
<namePart type="family">Post</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arul</namePart>
<namePart type="family">Menezes</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>Findings of the Association for Computational Linguistics: EMNLP 2022</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yoav</namePart>
<namePart type="family">Goldberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zornitsa</namePart>
<namePart type="family">Kozareva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</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</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Traditional machine translation (MT) metrics provide an average measure of translation quality that is insensitive to the long tail of behavioral problems. Examples include translation of numbers, physical units, dropped content and hallucinations. These errors, which occur rarely and unpredictably in Neural Machine Translation (NMT), greatly undermine the reliability of state-of-the-art MT systems. Consequently, it is important to have visibility into these problems during model development. Towards this end, we introduce SALTED, a specifications-based framework for behavioral testing of NMT models. At the core of our approach is the use of high-precision detectors that flag errors (or alternatively, verify output correctness) between a source sentence and a system output. These detectors provide fine-grained measurements of long-tail errors, providing a trustworthy view of problems that were previously invisible. We demonstrate that such detectors could be used not just to identify salient long-tail errors in MT systems, but also for higher-recall filtering of the training data, fixing targeted errors with model fine-tuning in NMT and generating novel data for metamorphic testing to elicit further bugs in models.</abstract>
<identifier type="citekey">raunak-etal-2022-salted</identifier>
<identifier type="doi">10.18653/v1/2022.findings-emnlp.379</identifier>
<location>
<url>https://aclanthology.org/2022.findings-emnlp.379</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>5163</start>
<end>5179</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SALTED: A Framework for SAlient Long-tail Translation Error Detection
%A Raunak, Vikas
%A Post, Matt
%A Menezes, Arul
%Y Goldberg, Yoav
%Y Kozareva, Zornitsa
%Y Zhang, Yue
%S Findings of the Association for Computational Linguistics: EMNLP 2022
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F raunak-etal-2022-salted
%X Traditional machine translation (MT) metrics provide an average measure of translation quality that is insensitive to the long tail of behavioral problems. Examples include translation of numbers, physical units, dropped content and hallucinations. These errors, which occur rarely and unpredictably in Neural Machine Translation (NMT), greatly undermine the reliability of state-of-the-art MT systems. Consequently, it is important to have visibility into these problems during model development. Towards this end, we introduce SALTED, a specifications-based framework for behavioral testing of NMT models. At the core of our approach is the use of high-precision detectors that flag errors (or alternatively, verify output correctness) between a source sentence and a system output. These detectors provide fine-grained measurements of long-tail errors, providing a trustworthy view of problems that were previously invisible. We demonstrate that such detectors could be used not just to identify salient long-tail errors in MT systems, but also for higher-recall filtering of the training data, fixing targeted errors with model fine-tuning in NMT and generating novel data for metamorphic testing to elicit further bugs in models.
%R 10.18653/v1/2022.findings-emnlp.379
%U https://aclanthology.org/2022.findings-emnlp.379
%U https://doi.org/10.18653/v1/2022.findings-emnlp.379
%P 5163-5179
Markdown (Informal)
[SALTED: A Framework for SAlient Long-tail Translation Error Detection](https://aclanthology.org/2022.findings-emnlp.379) (Raunak et al., Findings 2022)
ACL