This advanced course investigates the problems and techniques of natural language generation (NLG). Students learn about NLG tasks and architectures, both for text generation and for dialogue response generation. Several approaches are studied and compared, including template-based generation (with canned text and with syntactic templates), grammar-based generation (with CFGs and with systemic grammars), and XML-based generation. The course describes mainly English language generation, but students can do practical work on generation of Finnish or other languages.
The course is taught in English. The course consists of lectures and practical lab sessions.
The course assumes previous knowledge of language technology, including
Please contact me if you are not sure about meeting the prerequisites.
Time: Tuesdays 14-16, 14.01.03 - 15.02.03.
Place: Unix lab, Siltavuorenpenger 20A.
The course textbook is Reiter & Dale,
Building Natural Language Generation Systems,
Cambridge University Press, 2000.