Some project ideas.


- Self-training for LingPipe tagger
- Test LingPipe POS tagger on blog posts
- Implement Brill's part-of-speech tagger
- Features for chart parser
- Better edge hypothesizers for chart parser
- Generalize chart parser to use statistics
- Test named entity extractors on blog posts
- Implement a (partial) bibliographic matching system for 
  using LingPipe or similar. See http://www.cs.umd.edu/~getoor/Publications/icml03-ws.pdf and http://www.springerlink.com/content/r1723134248214t0/fulltext.pdf (accessible from OSU only)
- anything else you can get me to agree to by Friday 8 May.

Groups of two or three are recommended. Groups of one accepted only if
you give me an excellent reason