XTAG meeting, Thurs. May 10, 10:30, fishbowl

This week Anoop will present the following paper:

A Novel Use of Statistical Parsing to Extract Information from Text. S.
Miller, H. Fox, L. Ramshaw and R. Weischedel. Proc. of 1st Meeting of
NAACL. Apr 29-May 4, 2000, Seattle, WA.

Here is a short abstract:

The motivation is to view their approach (in a suitably modified form)
as a possible methodology for augmenting the XTAG system to output a
semantics associated with derivation trees.

We need some (semi) automatic way of assigning the individual semantic
building blocks to each elementary tree. The parser will combine these
blocks to give an output semantics for an input sentence.

Ideally we need a method that will work independent of whether we have
multiple parses or a single hand-corrected parse per sentence; or
whether the TAG grammar is the present XTAG grammar, a future modified
version of the XTAG grammar or even a radically different Treebank TAG