A Game Theory Approach for Multi-document Summarization

In today’s era, information has been growing exponentially on the web, due to which extraction of relevant and concise information has become a challenging task

Amreen Ahmad

2018

Scholarcy highlights

  • In today’s era, information has been growing exponentially on the web, due to which extraction of relevant and concise information has become a challenging task
  • To overcome the above problem, a fundamental tool known as summarization techniques has been used for understanding and organizing such large datasets
  • Researchers have been devoting a lot of effort to develop semantics-based models, so as to improve summarization performance
  • A versatile and principled game theory-based multi-document summarization framework integrated with Wikipedia ontology is proposed
  • The framework exploits the submodularity hidden in underlying ontology and is optimized using the proposed improved algorithm, to enhance the summarization performance
  • Results of the proposed approach were evaluated with the ROUGE evaluation metric for different multi-document summarization tasks against human-generated summaries and it outperformed DUC, TAC competitors, and other competitive methods

Need more features? Save interactive summary cards to your Scholarcy Library.