conclusion.tex (555B)
1 \begin{frame}{Conclusion} 2 \begin{block}{Take-home Message} 3 Topological information can be leverage for unsupervised relation extraction. 4 \end{block} 5 \begin{block}{Contributions} 6 \begin{itemize} 7 \item Explicitly modeled the aggregate setup for the unsupervised problem. 8 \item Provided proof on the quality of topological information. 9 \item Proposed an approach to exploit the mutual information between topological and linguistic features. 10 \end{itemize} 11 \end{block} 12 13 \medskip 14 15 Several directions still need to be explored. 16 \end{frame}