Conclusion.tex

\section{Conclusion}
\label{sec:Conclusion}

We proposed a novel problem of generating a digest that present a holistic 
view on a given Wikipedia event. We proposed a novel divergence-based 
framework for extracting excerpts from pseudo-relevant input news articles 
such that the global divergence between the digest and the query is minimized. 
As our method, we formulated the problem as an ILP for global inference to 
maximizes the overall relevance of the digest to the input event query while  
reducing inter-excerpt redundancies in our four dimensions of an event namely, text, time, gelocations, and entity. We compared several methods that 
were either derived from the literature or proposed by us, and found that the 
best method for generating an event digest is the one that adopts our 
divergence-based framework and considers all dimensions of an event.