Transductive Learning over Automatically Detected Themes for Multi-Document Summarization

Massih-Reza Amini(2), Nicolas Usunier(1)
(1) Laboratoire d'Informatique Paris 6              (2) National Research Council Canada
              4, Place de Jussieu                               123, boulevard Alexandre Taché         
75252 Paris 05 cedex                                  Gatineau, Canada         

We propose a new mthod for query-biased multi-document summarization, based on sentence extraction. The summary of multiple documents is created in two steps. Sentences are first clustered; where each cluster corresponds to one of the main themes present in the collection. Inside each theme, sentences are then ranked using a transductive learning to rank algorithm based on RankNet, in order to better identify those which are relevant to the query. The final summary contains the top-ranked sentences of each theme. Our approach is validated on DUC 2006 and DUC 2007 datasets.