Uncertainty-Aware Machine Translation Evaluation
This addresses the need for more trustworthy machine translation evaluation for researchers and practitioners, though it is incremental as it builds on existing COMET framework with uncertainty estimation methods.
The paper tackles the problem of unreliable quality predictions in neural machine translation evaluation metrics by introducing uncertainty-aware evaluation that provides confidence intervals alongside quality scores. The results show the method's performance across multiple language pairs and its usefulness in flagging critical translation mistakes.
Several neural-based metrics have been recently proposed to evaluate machine translation quality. However, all of them resort to point estimates, which provide limited information at segment level. This is made worse as they are trained on noisy, biased and scarce human judgements, often resulting in unreliable quality predictions. In this paper, we introduce uncertainty-aware MT evaluation and analyze the trustworthiness of the predicted quality. We combine the COMET framework with two uncertainty estimation methods, Monte Carlo dropout and deep ensembles, to obtain quality scores along with confidence intervals. We compare the performance of our uncertainty-aware MT evaluation methods across multiple language pairs from the QT21 dataset and the WMT20 metrics task, augmented with MQM annotations. We experiment with varying numbers of references and further discuss the usefulness of uncertainty-aware quality estimation (without references) to flag possibly critical translation mistakes.