Tangled up in BLEU: Reevaluating the Evaluation of Automatic Machine Translation Evaluation Metrics
Year: 2020
Type: conference-paper
Abstract: Automatic metrics are fundamental for the development and evaluation of machine translation systems.Judging whether, and to what extent, automatic metrics concur with the gold standard of human evaluation is not a straightforward problem.We show that current methods for judging metrics are highly sensitive to the translations used for assessment, particularly the presence of outliers, which often leads to falsely confident conclusions about a metric's efficacy.Finally, we turn to pairwise system ranking, developing a method for thresholding performance improvement under an automatic metric against human judgements, which allows quantification of type I versus type II errors incurred, i.e., insignificant human differences in system quality that are accepted, and significant human differences that are rejected.Together, these findings suggest improvements to the protocols for metric evaluation and system performance evaluation in machine translation.
Authors: Nitika Mathur, Timothy Baldwin, Trevor Cohn
Institution: The University of Melbourne
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Funder: Australian Research Council
OpenAlex ID: W3034269545