BLEU Meets COMET: Combining Lexical and Neural Metrics Towards Robust Machine Translation Evaluation
This work addresses the need for more reliable evaluation metrics in machine translation, particularly for detecting critical errors, though it is incremental as it builds on existing methods.
The paper tackles the problem of unreliable detection of critical errors like entity and number deviations in neural machine translation evaluation metrics by combining them with traditional lexical metrics. The result is improved robustness, with gains in correlation with human judgments and performance on challenge sets across multiple language pairs.
Although neural-based machine translation evaluation metrics, such as COMET or BLEURT, have achieved strong correlations with human judgements, they are sometimes unreliable in detecting certain phenomena that can be considered as critical errors, such as deviations in entities and numbers. In contrast, traditional evaluation metrics, such as BLEU or chrF, which measure lexical or character overlap between translation hypotheses and human references, have lower correlations with human judgements but are sensitive to such deviations. In this paper, we investigate several ways of combining the two approaches in order to increase robustness of state-of-the-art evaluation methods to translations with critical errors. We show that by using additional information during training, such as sentence-level features and word-level tags, the trained metrics improve their capability to penalize translations with specific troublesome phenomena, which leads to gains in correlation with human judgments and on recent challenge sets on several language pairs.