Feature-Rich Discriminative Phrase Rescoring for SMT

This paper proposes a new approach to phrase rescoring for statistical machine translation (SMT). A set of novel features capturing the translingual equivalence between a source and a target phrase pair are introduced. These features are combined with linear regression model and neural network to predict the quality score of the phrase translation pair. These phrase scores are used to discriminatively rescore the baseline MT system’s phrase library: boost good phrase translations while prune bad ones. This approach not only significantly improves machine translation quality, but also reduces the model size by a considerable margin.