Key points are not available for this paper at this time.
Automatic abstracting is a challenging task for natural language processing. It depends not only on sentence interpretation and the local context representation this requires, but also on the recognition and use of large-scale discourse structure. This paper describes research investigating the nature of different approaches to discourse representation and their value for summarising. This work is focussed on comparative analysis, illustrated in the paper through the provision of different forms of representation, and different strategies for summary formation, for a short example text.
Karen Spärck Jones (Mon,) studied this question.