Welcome to the META-NET course on advanced MT resources, Prague, Dec 16-17, 2010
Nov. 16, 2010: Program and registration are up.
Oct. 15, 2010: Website created.
About the course
META-NET is a Network of Excellence dedicated to building the technological foundations of a multilingual European information society.
META-NET organizes various events to promote its goals, among them technology shows (e.g. the META-FORUM event in Brussels) and training events like this one.
In this course, we will present resources being created or currently available for adding semantics to machine translation systems. Current machine translation systems both in the research stage and in commercial production use statistical methods that learn from parallel corpora. The most successful ones, such as the open-source MOSES-based system can work on the phrase level; however, none of them use semantic relations present in the text explicitly. It is believed that this might be important especially for languages for which limited parallel corpora exist. The course will present language resources, both parallel as well as monolingual, which will provide the basis for semantic analyzers that could possibly be used in future MT systems.
This two-day intensive course is aimed at researchers and graduate students working in or wanting to join the area of Machine translation research and development.
The course is free of charge, but registration is required. All participants are also invited (also free of charge) to attend the CLARA treebank(ing) course, which immediately precedes this META-NET training event (in fact, these two education and training events share the lectures on Thursday, Dec. 16, the day when the lectures will take place in the historical 17th century refectory of the Computer Science School building in Prague).
For the venue, hotel suggestions, maps, visa and other information, please go to the "Venue" link on the left.