Annually, at the Conference of Machine Translation (WMT), the Metrics Shared Task organizers conduct the Machine Translation (MT) meta-evaluation, assessing MT metrics' capabilities according to their correlation with human judgments. Their results guide researchers toward enhancing the next generation of metrics and MT systems. With the recent introduction of neural metrics, the field has witnessed notable advancements. Nevertheless, the inherent opacity of these metrics has posed substantial challenges to the meta-evaluation process. This work highlights two issues with the meta-evaluation framework currently employed in WMT, and assesses their impact on the metrics rankings. To do this, we introduce the concept of sentinel metrics, which are designed explicitly to scrutinize the meta-evaluation process's accuracy, robustness, and fairness. By employing sentinel metrics, we aim to validate our findings, and shed light on and monitor the potential biases or inconsistencies in the rankings. We discover that the present meta-evaluation framework favors two categories of metrics: i) those explicitly trained to mimic human quality assessments, and ii) continuous metrics. Finally, we raise concerns regarding the evaluation capabilities of state-of-the-art metrics, emphasizing that they might be basing their assessments on spurious correlations found in their training data.
Guardians of the Machine Translation Meta-Evaluation: Sentinel Metrics Fall In! / Perrella, Stefano; Proietti, Lorenzo; Scire', Alessandro; Barba, Edoardo; Navigli, Roberto. - 1:(2024), pp. 16216-16244. (Intervento presentato al convegno Association for Computational Linguistics tenutosi a Bangkok; Thailand) [10.18653/v1/2024.acl-long.856].
Guardians of the Machine Translation Meta-Evaluation: Sentinel Metrics Fall In!
Stefano Perrella
Co-primo
;Lorenzo Proietti
Co-primo
;Alessandro Scire'
Secondo
;Edoardo Barba
Penultimo
;Roberto Navigli
Ultimo
2024
Abstract
Annually, at the Conference of Machine Translation (WMT), the Metrics Shared Task organizers conduct the Machine Translation (MT) meta-evaluation, assessing MT metrics' capabilities according to their correlation with human judgments. Their results guide researchers toward enhancing the next generation of metrics and MT systems. With the recent introduction of neural metrics, the field has witnessed notable advancements. Nevertheless, the inherent opacity of these metrics has posed substantial challenges to the meta-evaluation process. This work highlights two issues with the meta-evaluation framework currently employed in WMT, and assesses their impact on the metrics rankings. To do this, we introduce the concept of sentinel metrics, which are designed explicitly to scrutinize the meta-evaluation process's accuracy, robustness, and fairness. By employing sentinel metrics, we aim to validate our findings, and shed light on and monitor the potential biases or inconsistencies in the rankings. We discover that the present meta-evaluation framework favors two categories of metrics: i) those explicitly trained to mimic human quality assessments, and ii) continuous metrics. Finally, we raise concerns regarding the evaluation capabilities of state-of-the-art metrics, emphasizing that they might be basing their assessments on spurious correlations found in their training data.File | Dimensione | Formato | |
---|---|---|---|
Perrella_Guardians_2024.pdf
accesso aperto
Note: PDF: https://aclanthology.org/2024.acl-long.856.pdf
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
2.58 MB
Formato
Adobe PDF
|
2.58 MB | Adobe PDF |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.