While much work has been done in the area of extractive summarization, there has been limited study in abstractive summarization as this is much harder to achieve (going by the definition of true abstraction). Existing work in abstractive summarization may not be truly abstractive, and even if it is, it may not be fully automated. This page contains a very small collection of summarization methods that are non-extractive. Please take note that this is by no means an exhaustive list, it would just be a starting point.
- UNL Document Summarization
Virach S, Potipiti T and Charoenporn T. UNL document summarization. Proceedings of the first in-ternational workshop on multimedia annotation (MMA2001), Tokyo, Japan, January 2001
DeJong G. An Overview of the FRUMP system In Strategies for natural language processing, W. G. Lehnert and M. H. Ringle (eds.), 149–176. Hillsdale, New Jersey: Erlbaum 1982
Radev D R and McKeown K R. Generating natural language summaries from multiple on-line sources. Computational Linguistics 1998; 24(3):469–500
Summary: Involves shallow syntactic and semantic analysis, concept identification, and text regeneration. Method was developed through the study of a corpus of abstracts written by professional abstractors.
Harabagiu, S., Lacatusu, F. Generating Single and Multi-Document Summaries with GISTEXTER. In Workshop on Text Summarization (In Conjunction with the ACL 2002 and including the DARPA/NIST sponsored DUC 2002 Meeting on Text Summarization) Philadelphia, USA, 2002.
- SUMUM Generating Indicative-Informative Summaries with SumUM (Horacio Saggion , Guy Lapalm)
Ganesan, K. A., C. X. Zhai, and J. Han, “Opinosis: A Graph Based Approach to Abstractive Summarization of Highly Redundant Opinions”, Proceedings of the 23rd International Conference on Computational Linguistics (COLING), Beijing, China, pp. 340–348, 2010.
Summary: Uses a word graph data structure for each input document to find promising paths that act as candidate summaries. Leverages 3 key properties of the graph that helps generate summaries that are more abstractive (shallow abstraction). This method is syntax lean – uses only POS tags.