SUMMARIST Multilingual Text Summarization

SUMMARIST is a multilingual text summarization system developed in the Natural Language Group of the Information Sciences Institute of the University of Southern California.

It produces extract-type summaries (verbatim extractions of sentences and keywords) from newspaper, newswire, and other texts. It operates on texts in English, Spanish, French, German, Indonesian; older versions have also operated on Arabic and Japanese.

In this demonstration, SUMMARIST is coupled with the machine translation system SYSTRAN, developed by Systran Software Inc. of La Jolla, CA. This is the same engine used by AltaVista and called Babelfish. In this joint demonstration you see the power of summarizing and then translating only the summary (or of summarization or translation, separately).

Running SUMMARIST and SYSTRAN

To use the system, point your browser at http://128.9.208.230/must/index.html

Perform the following steps:

1. Determine the URL of the webpage text you wish to summarize and/or translate. You can find news articles by selecting one of the news stories in the following pages and copying the URL:
  • French: www.lemonde.fr
  • Spanish: www.elpais.es
  • German: www.spiegel.de
  • English: abcnews.go.com

2. Type the URL into the main window

3. Select the length of summary you prefer

4. Select the language translation direction

5. Click on the appropriate button:

  • Summarize and Translate -- run both SUMMARIST and Babelfish
  • Summarize -- run ISI's SUMMARIST only
  • Translate -- run SYSTRAN's Babelfish only

The result should return on your screen in under three seconds; if it takes longer, something is wrong. You will see:

  • the original language summary
  • the translation of the summary
  • a list of key words and phrases characterizing the text, produced by SUMMARIST

If you wish, you can also see the paragraphs surrounding the sentences extracted by SUMMARIST, in order to get a better feel for the context.

For information, please contact:

Dr. Eduard Hovy or Dr. Chin-Yew Lin
Information Sciences Institute of USC
hovy@isi.edu or cyl@isi.edu
http://www.isi.edu/natural-language/nlp-at-isi.html

[ NLG overview | Project Members | Publications | Projects | Demonstrations ]