ExampleBased Machine Translation English to Chinese

1
Example-Based Machine Translation (English to
Chinese)
Presented by Jackson Lee Supervised by Dr.
Xue Li
Entering the URL of the English website and
submit for the Chinese Equivalent web Page
Translation

• Julie made a chocolate cake
• ???????????

John ate an Apple ????????

John ate a chocolate cake???????????
Figure 2 The equivalent Chinese web page
translated from English
Figure 1 The original web page in English
SIGNIFICANCE OF THE PROJECT

There are 80 percent of web
pages in English and Chinese is the 2nd most used
language in the world which makes It meaningful
to be able to translate web pages in English into
equivalent Chinese web pages. The project aims
to develop a machine translation web service that
can directly solve that problem for Chinese users
with difficulties in English. With the new
feedback function the Example based MT, the
quality of the translation on internet can be
greatly improved. My contributions toward this
project include the XML database for the Chinese
translated units and the user interacting
feedback function.

• RESEARCH ISSUE
• Efficiency and accuracy of MT techniques
• Self-learning process
• Feed-back function currently used on web
  translation
• Designed a web page for accessing the web
  service
• Server extension - to retrieve and parse HTML
• content in order to extract text for
  translation purpose
• Decomposing English text into smaller single
• sentences for translation
• Database storage - appropriate Chinese
  translation
• for each input sentence

Figure 3 The whole translation process
flowchart.
EXAMPLE-BASSED MACHINE TRANSLATION

The
English-Chinese translation is done using the
example-based approach translating sentence by
sentence. This approach requires large amount of
translated sentences already stored in the
database. The system retrieves the web page in
English then retrieves all the text from the html
tags. These sentences are then divided into
smaller units then the service finds the best
suits translation from the database. These
translated sentences then are recomposed and
combined and set out in the html format. This is
a continuing project followed on from a Master
Student, Hung Phanvan supervised by Dr. Xue Li.
His work is adapted by Jimmy Kung, Jackson Lee
and Flora Yeh. Currently there are several
existing web translations available on the world
wide web. Altavistas BabelFish. WorldLingo web
Translation, InterTran and various other
softwares do similar jobs.