A Discourse Parser Language Model Based on Improved Neural Network in Machine Translation

摘要

The development of statistical machine translation technology is so fast and there are many new models and methods. This obtains great achievement in translation of simple sentence or fixed sentence with certain applications. However, there still exists poor coherence and low readability complex long sentence translation. In order to measure discourse coherence and adapt to more tasks more effectively, this paper puts forward a word vector characteristic-based the improved method of recurrent neural network language model. This method increases feature layer in input layer. The improved model structure adds context word vector through feature layer during model training and enhances learning ability in long-distance information restriction. The experiment result shows that our proposed hierarchical recurrence neural network-based discourse language model has a better performance which is beyond current optimal system.