Improving of Hierarchical Phrase-based SMT Model Based on Prosodic Structure Information
Yidong Chen · Computer Knowledge and Technology · 2013
With the intensive study of natural language processing,it does not only need to tap more potential linguistic information,but also need to try predict these knowledge by machine learning.And in order to improve the performance of machine translation,as well as for the computational linguistics..Prosody,as the knowledge of phonetics,has been widely applied to speech synthesis,text-to-speech,voice translation and other aspects of phonetics,which has played an irreplaceable role in speech processing.we introduce how and in which way prosodic information improve machine translation.And it proves that prosodic information is useful for hierarchical-phrased model in the experiment.