Optimizing Non-Decomposable Evaluation Metrics for Neural Machine Translation

While optimizing model parameters with respect to evaluation metrics has recently proven to benefit end to-end neural machine translation (NMT), the evaluation metrics used in the training are restricted to be defined at the sentence level to facilitate online learning algorithms. This is undesirable because the final evaluation metrics used in the testing phase are usually non-decomposable (i.e., they are defined at the corpus level and cannot be expressed as the sum of sentence-level metrics). To minimize the discrepancy between the training and the testing, we propose to extend the minimum risk training (MRT) algorithm to take non-decomposable corpus-level evaluation metrics into consideration while still keeping the advantages of online training. This can be done by calculating corpus-level evaluation metrics on a subset of training data at each step in online training. Experiments on Chinese-English and English-French translation show that our approach improves the correlation between training and testing and significantly outperforms the MRT algorithm using decomposable evaluation metrics.