NYU Proteus Project

1
Prof. Ralph Grishman Director, Proteus
Project Research in Natural Language Processing
NYU Proteus Project
Information Extraction finding information in
text about specified types of events and
populating a data base

• George Garrick, 40 years old, president of the
  London-based European Information Services Inc.,
  was appointed chief executive officer ofNielsen
  Marketing Research, USA.

2
The challenge of Information Extraction
finding the many ways an event may be
described appointed George Garrick was named
to the position of CEO selected as the
next will join the firm as Our group is
developing strategies to learn these different
forms automatically.
Web site nlp.cs.nyu.edu
Approaches supervised learning annotate a
lot of examples and train a model semi-supervised
learning annotate a few examples and find
more examples automatically from the Web (rely
on complementary features) active
learning select examples to ask the user about
minimize the information the user must provide