poster

1 files changed, 22 insertions, 46 deletions

diff --git a/poster/Finale_poster/poster.tex b/poster/Finale_poster/poster.tex
index aef5490..60c1bb5 100644
--- a/poster/Finale_poster/poster.tex
+++ b/poster/Finale_poster/poster.tex
@@ -26,20 +26,10 @@
 \usepackage{graphicx}
 \usepackage{booktabs}
 
-%----------------------------------------------------------------------------------------
-%	TITLE SECTION
-%----------------------------------------------------------------------------------------
-
 \title{Bayesian and Active Learning for Graph Inference} % Poster title
-
 \author{Thibaut Horel, Jean Pouget-Abadie} % Author(s)
 
-%\institute{Harvard University} % Institution(s)
-%----------------------------------------------------------------------------------------
 \begin{document}
-\addtobeamertemplate{block end}{}{\vspace*{2ex}} % White space under blocks
-\addtobeamertemplate{block alerted end}{}{\vspace*{2ex}} % White space under
-
 \setlength{\belowcaptionskip}{2ex} % White space under figures
 \setlength\belowdisplayshortskip{2ex} % White space under equations
 
@@ -48,11 +38,6 @@
 \begin{column}{\sepwid}\end{column}
 \begin{column}{\onecolwid} % The first column
 
-%----------------------------------------------------------------------------------------
-%	INTODUCTION
-%----------------------------------------------------------------------------------------
-
-
 \begin{block}{Problem}
   \emph{How to recover an unknown network from the observation of contagion
   cascades?}
@@ -81,9 +66,9 @@
   \vspace{1em}
 \begin{figure}
   \centering
-  \includegraphics[scale=2]{drawing.pdf}
 \end{figure}
 \end{block}
+  \includegraphics[scale=1.5]{drawing.pdf}
 
 \begin{block}{MLE}
   \begin{itemize}
@@ -123,20 +108,33 @@
   source node?}
 \end{block}
 
-\begin{block}{Heuristic 1 (Baseline)}
+\begin{block}{Heuristic 1}
+  \begin{itemize}
+    \item Choose source proportional to estimated degree
+    \item Intuition:
+      \begin{itemize}
+        \item deeper cascades $\implies$ more data
+        \item easier to learn non-edges than edges. Higher degree $\implies$
+          more edge realizations at the same price.
+      \end{itemize}
+  \end{itemize}
 \end{block}
 
 \begin{block}{Heuristic 2}
   \begin{itemize}
-    \item Choose source to maximize mutual information
+    \item Choose source proportional to mutual information
     \begin{equation*}
-      \begin{split}
-        I(\bx ,\Theta | x^0 = i) = &
-      E_{\substack{\Theta \sim D_t \\ x |
-      \Theta, i \sim D'}}\log p(x|\Theta) \\ - & E_{\substack{\Theta
-      \sim D_t \\ x' | \Theta, i \sim D'}} \log p(x')
-    \end{split}
+      I((X_t) ,\Theta | x^0 = i) = - H(\Theta | (X_t), X_0 = i) + H(\Theta)
     \end{equation*}
+    \item Exact strategy requires knowing true distribution of $(X_t)$
+    \item Use estimated $\Theta$ to compute $H(\Theta | (X_t), X_0 = i)$
+\end{itemize}
+\end{block}
+
+\begin{block}{Heuristic 3}
+  \begin{itemize}
+    \item Choose source proportional to mutual information of first step of
+      cascade and $\Theta$:
   \end{itemize}
 \end{block}
 
@@ -146,26 +144,6 @@
 %-----------------------------------------------------------------------------
 
 \begin{column}{\onecolwid} % The third column
-  \begin{block}{Experimental validation}
-    \begin{figure}
-      \centering
-      \includegraphics[scale=1.2]{../images/watts_strogatz.pdf}
-    \caption{Watts-Strogatz, $300$ nodes, $4500$ edges, uniform edge weights,
-    constant $p_{init}$}
-    \end{figure}
-  \end{block}
-
-
-\begin{block}{Conclusion}
-  \begin{itemize}
-    \item Introduce Generalized Linear Casacade model
-    \item Better finite sample guarantees~\cite{Netrapalli:2012, Abrahao:13,
-        Daneshmand:2014}
-    \item Interpretable conditions on Hessian
-    \item Lower bound+approximately sparse case developed in full
-      paper~\cite{Pouget:2015}
-  \end{itemize}
-\end{block}
 
 %-----------------------------------------------------------------------------
 %	REFERENCES
@@ -185,6 +163,4 @@
 
 \end{frame} % End of the enclosing frame
 
-
-
 \end{document}