Machine Translation And Human Intervention

With the rapid growth of global economy, traditional human translation (HT) is far from being able to meet the increasing translation demand. Thus, Machine Translation (MT) well known for its high speed, high consistency, and low cost becomes the worldwide hot topic and has been under great discussion. However, in spite of fifty years of development, MT hasn't got any real break-through in improving its output quality and repeatedly makes errors of this kind or that which human translators seldom make. One of the main reasons is that machine lacks the real world knowledge characteristic of human beings. In view of this, in this thesis the author presents some superficial ideas on how to improve MT output quality based on the experiments on Chinese/English MT systems.This thesis traces back the MT history in the world and in China, makes some introductions about MT approaches, systems and evaluation, and lists some common Chinese/English MT software and online MT websites. Meanwhile, this thesis makes a clear analysis of various errors made by Chinese/English MT systems, puts forward methods of human intervention-pre-editing and post-editing-to improve MT quality and shorten the gap between HT and MT. Last but not the least, the thesis predicts the future development of MT, proposing perfecting post-editing service, Artificial Intelligence (AI), and Corpus to improve MT quality.