Title:	 Corpus-Based Methods in Chinese Morphology

Speaker: Dr. Richard Sproat 
         AT&T Labs -- Research 180 Park Ave, Room B207
	 Florham Park, NJ 07932, USA +1-973-360-8490
	 rws@research.att.com

Summary: The course will cover selected topics in corpus-based
         linguistics and statistical natural language processing. I
         will show how some of these techniques can be applied to
         problems related more specifically to Chinese: specific topic
         areas will include statistical analysis of morphological
         productivity in modern Chinese, morphological phenomena such
         as abbreviation (suoxie), questions of morphological
         structure in earlier forms of Chinese and more practical NLP
         problems such as Chinese word segmentation and named entity
	 recognition.

Tutorial Outline: 3 hour tutorial, assuming 15 minute break in the middle.
                  This is tentative and will likely change as I
                  develop materials...

	 0:00 - 0:30   Issues in Chinese morphology
	 0:30 - 1:30   Measures of association and measures of productivity
	 1:30 - 1:45   Break
	 1:45 - 2:00   Productivity in Chinese morphology: Root compounds
	 2:00 - 2:30   Chinese word segmentation
	 2:30 - 2:45   Theories of Suoxie
	 2:45 - 3:00   Measures of association and concepts of ancient
		       Chinese morphology

Bio:	 Richard Sproat has worked in many areas of speech and
         computational linguistics including computational morphology,
         corpus-based methods, text-to-speech synthesis, the
         computational analysis of writing systems, and text-to-scene
         conversion; as well as a number of areas of theoretical
         linguistics. He has also done research in Chinese
         linguistics, both theoretical work on Chinese morphology, and
         more practical work on Chinese word segmentation and
         text-to-speech synthesis.