Statistical Machine Translation Features with Multitask Tensor Networks
This work addresses translation quality for SMT systems, offering incremental improvements through complementary methods.
The paper tackled improving Statistical Machine Translation by proposing new neural network features, tensor layers for higher-order interactions, and multitask learning, resulting in gains of +2.7 and +1.8 BLEU points for Arabic-English and Chinese-English translation over a state-of-the-art system.
We present a three-pronged approach to improving Statistical Machine Translation (SMT), building on recent success in the application of neural networks to SMT. First, we propose new features based on neural networks to model various non-local translation phenomena. Second, we augment the architecture of the neural network with tensor layers that capture important higher-order interaction among the network units. Third, we apply multitask learning to estimate the neural network parameters jointly. Each of our proposed methods results in significant improvements that are complementary. The overall improvement is +2.7 and +1.8 BLEU points for Arabic-English and Chinese-English translation over a state-of-the-art system that already includes neural network features.