diff --git a/bilder/Ranking.png b/bilder/Ranking.png
new file mode 100644
index 0000000000000000000000000000000000000000..4af4ebbed8e3a6d2784d72015492d8359fcd3696
Binary files /dev/null and b/bilder/Ranking.png differ
diff --git a/slides/dataset.tex b/slides/dataset.tex
index 395d9f5383864001e03be6a48f6db82fc59d86a6..35d112b33825bb743f45bd444b36f6c63b1b47a1 100644
--- a/slides/dataset.tex
+++ b/slides/dataset.tex
@@ -3,51 +3,34 @@
 \subsection{Corpus}
 \begin{frame}
 	\frametitle{Corpus}
-	\begin{itemize}
-		\item For our study, we used the Webis-ArgRank2017 dataset by \cite{wachsmuth:2017a}.
-		\item Webis-ArgRank2017 contains a ground-truth argument graph
-		      as well as a benchmark for argument ranking
-		\item The data was originally collected within the Argument-Web
-	\end{itemize}
-	\begin{figure}
-		\includegraphics[width=0.4\linewidth]{bilder/DatasetGlobalView2.png}
-		\caption{Complete Webis-ArgRank2017.}
-	\end{figure}
-\end{frame}
-
-\begin{frame}
-	\frametitle{Corpus}
-	\textbf{The argument graph} $G = (A, E)$ \textbf{ is a directed graph with:}
-	\begin{itemize}
-		\item Each argument-node $a_i \in A$ consisting of a conclusion $c_i$ and a not-empty set of premises $P_i$ $\Rightarrow$ $a_i = \langle c_i, P_i \rangle$
-		\item An edge $(a_j, a_i)$ is given if the conclusion $a_j$ is used as a premise of $a_i$
-		\item Consequently, $P_i = \{c_1,...,c_k\}, k \geq 1$
-	\end{itemize}
-	\begin{figure}
-		\includegraphics[width=0.7\linewidth]{bilder/DatasetLocalView2.png}
-		\caption{Part of the Argument-Web used to create Webis-ArgRank2017.}
-	\end{figure}
+	\begin{columns}
+		\column{0.5\textwidth}
+			\begin{itemize}
+				\item Webis-ArgRank2017 dataset by \cite{wachsmuth:2017a}.
+				\item Ground-truth argument graph
+				\item Benchmark for argument ranking
+			\end{itemize}
+		\column{0.5\textwidth}
+			\begin{figure}
+				\includegraphics[width=1.2\linewidth]{bilder/DatasetGlobalView2.png}
+			\end{figure}
+		\end{columns}
 \end{frame}
 
 \subsection{Benchmark Argument Ranking}
 \begin{frame}
 	\frametitle{Benchmark Argument Ranking}
-	\textbf{\cite{wachsmuth:2017a} kept those arguments fulfilling:}
-	\begin{itemize}
-		\item The conclusion was part in more than one argument
-		\item Real claim
-	\end{itemize}
-	\textbf{Additionally, an argument had to be:}
-	\begin{itemize}
-		\item a valid counterargument
-		\item based on reasonable premises
-		\item logically consistent
-	\end{itemize}
-	\begin{block}{In total}
-		110 argument with a total of 32 conclusions were ranked by 7 experts.
-		
-		Each annotated how much the premises contributed to the acceptance of the conclusion.
-
-		All experts agreed with Kendall's $\tau \approx 0.36$.
-	\end{block}
+	\begin{columns}
+		\column{0.5\textwidth}
+		\begin{itemize}
+			\item 110 arguments
+			\item 32 conclusions
+			\item 7 annotators
+			\item $\tau \approx 0.36$
+		\end{itemize}
+		\column{0.5\textwidth}
+			\begin{figure}
+				\includegraphics[width=0.6\linewidth]{bilder/Ranking.png}
+			\end{figure}
+		\end{columns}
 \end{frame}
\ No newline at end of file