Network Machine Translation, Information Flow Processing

The development of computer technology, especially INTERNET technology, has brought Machine Translation a broad application background and new problems as well. The practicability of an MT system requires adaptation and expansion. There are two highlights: various computer language information MT pre-processing and MT peripheral software design.Our work is mainly on "Huajian Network&Dynamic Machine Translation System" . From the two aspects listed above, the characteristics of our research are as follows:1. Based on the analysis of general computer documents, we bring forward a knowledge base mechanism based MT pre-processing method. We applied the idea to the HTML documents MT pre-processing design. The results indicate that knowledge base mechanism not only improves the robust and expansion of our system, but also enables some HTML tag knowledge acquisition. In addition, we discuss the computer documents linguistically processing and point out that the document format is an important knowledge source for disambiguation.2. We designed and developed "Huajian Network&Dynamic Machine Translation System" . Due to the text mirror buffer introduced, the dynamic MT subsystem exactly enables machine translation compared with the common electronic dictionaries. Our network MT subsystem can be connected to Netscape Navigator seamlessly and provides various functions. The comprehensive performance of our system excels other similar domestic ones.