Soft: new open source NLP software

Philippe Blache pb at
Tue Jun 20 12:51:48 UTC 2000

From: Jason Baldridge <jmb at>

Hi all,

This is an announcement to let people know of the availability of some
new open source NLP software.  Here are the descriptions and where you
can go to find out more.

The Quipu Grok Library - a library of Java components for performing
various natural language tasks.  These include several preprocessing
tasks (pos tagging, sentence detection, etc.), chart parsing, a large
categorial grammar for English (induced from the Penn treebank), and
some knowledge representation components (basic coreference, salience
tracking, etc.)  Homepage: <>.

The Quipu OpenNLP API - a preliminary collection of Java interfaces
that we are designing in conjunction with Grok in order to standardize
how NLP modules interact. Homepage: <>.

The Quipu Maximum Entropy Package - an Java implementation of the
maximum entropy framework.  It allows you to train, evaluate, and use
maxent models.  Homepage: <>.

"Quipu" is a moniker which we use to refer to several students at the
University of Edinburgh who are interested in building open source
natural language software.  This is not an exclusive club, and we
welcome anyone else who would like to contribute to this cause.
Homepage: <>.

The software described above is distributed under the GNU Public
Licenses (GPL and LGPL). This means you are free to download both the
binaries and the source code. You are also welcome to modify the
source code for your own applications.  There are certain restrictions
spelled out in the licenses, the main impact of which is to ensure
that modifications (e.g. improvements and bug fixes) and new software
which incorporates GPL'ed code is made available to the community.

Our hope is that NLP researchers don't have to keep re-inventing basic
NLP components every time they want to build a new system.  We hope to
achieve this through modularity and the open source philosophy.

Please send this announcement along to any mailing lists whose readers
you think would be interested.  Let us know if you have any questions!

Jason Baldridge: jmb at
Gann Bierner: gbierner at

Message diffusé par la liste Langage Naturel <LN at>
Informations, abonnement :
English version          :
Archives                 :

More information about the Ln mailing list