The XTAG Project
XTAG is an on-going project to develop a wide-coverage grammar for
English using a feature-based and lexicalized Tree Adjoining Grammar
formalism. XTAG also serves as an system for the development of TAGs
and consists of a predictive left-to-right parser, an X-windows
interface, a morphological analyzer, and a part-of-speech tagger.
Contents
A brief introduction to the software components of XTAG. The system is
currently implemented in Common Lisp and C. The graphical interface
runs under X windows.
There are two executable versions of XTAG currently installed: one for
SunOS 4.1.3 machines and the other for Solaris/SunOS 5.4 Sparc
machines. You can find which version to run by using
uname as follows:
% uname -sr
On Penn CIS computers the answer will either be SunOS 5.4
(the machine runs Solaris), or SunOS 4.1.3. Change the
directory to /mnt/linc/xtag/work/xtag/bin/ and check to
see if your DISPLAY environment variable is set. Then
follow the following instructions:
- For Solaris/SunOS 5.4: Run
Xtag.solaris.
- For SunOS 4.1.3: Run
Xtag.sunos.
For looking at the Xtag syntactic lexicon use xsyn which
is also resides in /mnt/linc/xtag/work/xtag/bin/. It is
used to examine database verions of the syntactic lexicon (e.g. the
English syntactic lexicon is
/mnt/linc/xtag/work/xtag/data/english/syntax/syntax.db).
Answers to some XTAG mysteries for new users:
- How do I insert special characters like alpha, beta, etc. in a
text input window?
In any text input window you can type the character for alpha, beta,
and epsilon by typing respectively Control+shift+a,
Control+shift+b, and Control+shift+e.
- How do I insert a subscript character in a node name?
Subscript characters can be entered in any text input window by
holding down the modifier key (the diamond key on Sun keyboards or the
Alt key on other keyboards depending on the keyboard
binding).
- The XTAG text input window misbehaves when I use it in
OpenWindows or the
olwm window manager.
In your ~/.Xdefaults or
~/.XTAG-defaults file set the following resource
definition:
OpenWindows.FocusLenience: True
A list of papers and technical reports that are available online. A
very incomplete list.
-
A Freely Available Wide Coverage Morphological Analyzer for English
Daniel Karp, Yves Schabes, Martin Zaidel and Dania Egedi. In
Proceedings of COLING '92.
-
A Freely Available Syntactic Lexicon for English
Dania Egedi and Patrick Martin. In the Proceedings of the
International Workshop of Sharable Natural Language Resources,
Nara, Japan, August 1994.
-
XTAG System - A Wide Coverage Grammar for English.
Christy Doran, Dania Egedi, Beth Ann Hockey, B. Srinivas, and
Martin Zaidel. In Proceedings of COLING '95.
-
A Lexicalized Tree Adjoining Grammar for English
The XTAG Research Group. Available as IRCS Technical Report
(IRCS-95-03).
-
Some Novel Applications of Explanation-Based Learning to Parsing
Lexicalized Tree-Adjoining Grammars
B. Srinivas, Aravind Joshi. In ACL '95.
[
Penn |
IRCS |
CIS Dept |
Ling Dept |
LDC
]
For further information about the XTAG project and software
availability send mail to
xtag-request@linc.cis.upenn.edu. Page maintained by Anoop Sarkar. Comments to
anoop@linc.cis.upenn.edu.
that
describes the wide coverage English grammar developed using XTAG
(postscript).