TY - GEN
T1 - Pushing the Right Buttons
T2 - Adversarial Evaluation of Quality Estimation
AU - Kanojia, Diptesh
AU - Fomicheva, Marina
AU - Ranasinghe, Tharindu
AU - Blain, Fred
AU - Orăsan, Constantin
AU - Specia, Lucia
N1 - Accepted to WMT 2021 Conference co-located with EMNLP 2021. 14 pages with a 4 page appendix
PY - 2021/9/22
Y1 - 2021/9/22
N2 - Current Machine Translation (MT) systems achieve very good results on a growing variety of language pairs and datasets. However, they are known to produce fluent translation outputs that can contain important meaning errors, thus undermining their reliability in practice. Quality Estimation (QE) is the task of automatically assessing the performance of MT systems at test time. Thus, in order to be useful, QE systems should be able to detect such errors. However, this ability is yet to be tested in the current evaluation practices, where QE systems are assessed only in terms of their correlation with human judgements. In this work, we bridge this gap by proposing a general methodology for adversarial testing of QE for MT. First, we show that despite a high correlation with human judgements achieved by the recent SOTA, certain types of meaning errors are still problematic for QE to detect. Second, we show that on average, the ability of a given model to discriminate between meaning-preserving and meaning-altering perturbations is predictive of its overall performance, thus potentially allowing for comparing QE systems without relying on manual quality annotation.
AB - Current Machine Translation (MT) systems achieve very good results on a growing variety of language pairs and datasets. However, they are known to produce fluent translation outputs that can contain important meaning errors, thus undermining their reliability in practice. Quality Estimation (QE) is the task of automatically assessing the performance of MT systems at test time. Thus, in order to be useful, QE systems should be able to detect such errors. However, this ability is yet to be tested in the current evaluation practices, where QE systems are assessed only in terms of their correlation with human judgements. In this work, we bridge this gap by proposing a general methodology for adversarial testing of QE for MT. First, we show that despite a high correlation with human judgements achieved by the recent SOTA, certain types of meaning errors are still problematic for QE to detect. Second, we show that on average, the ability of a given model to discriminate between meaning-preserving and meaning-altering perturbations is predictive of its overall performance, thus potentially allowing for comparing QE systems without relying on manual quality annotation.
KW - cs.CL
KW - cs.AI
UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85127214482&origin=inward&txGid=8536a111018f6c9e929be4bd5e64edfc
M3 - Conference contribution
SN - 978-1-954085-94-7
SP - 625
EP - 638
BT - Proceedings of the Sixth Conference on Machine Translation
PB - Association for Computational Linguistics (ACL)
ER -