Introduction

Automatic text summarisation consists of condensing the most relevant information from a given text. The techniques used may vary greatly depending on the compression rate, the function of the summary (informative, indicative...), the genre of the text, the language (or languages) of the original texts, the number of documents summarised each time (mono- or multi-document), etc.

Concerning the research performed by us in this area, we may cite:

• Generation of extracts by selecting sentences, using genetic algorithm. This has been applied to an on-line adaptive information system for university students, in order to condense the full contents of web sites according to their temporal needs.
• Generation of headlines from newswire articles, result of our participation in the Document Understanding Conferences 2003 and 2004.
• Generation of summaries of educational material obtained from the Internet.
• Generation of multi-document summaries, in answer to a question provided by the user, result of our participation in DUC-2005.

Publications

On-line demos

Some external links

• Document Understanding Conferences.