1 files changed, 27 insertions, 23 deletions
diff --git a/aied2018/presentation/motivation.tex b/aied2018/presentation/motivation.tex
index 69db435..6d6b95f 100644
--- a/aied2018/presentation/motivation.tex
+++ b/aied2018/presentation/motivation.tex
@@ -1,27 +1,31 @@
-	What is wrong with the following Python program that prints all divisors?
-			\begin{columns}
-			\begin{column}{0.50\textwidth}
-\begin{Verbatim}
+What is wrong with the following program that prints all divisors?
+
+\vspace{0.5cm}
+
+\begin{Verbatim}[xleftmargin=6mm]
 \textbf{def} divisors(n):
-  \textbf{for} d \textbf{in} range(1, \red{n}):
+  \textbf{for} d \textbf{in} range(1,\red{n}):
     \textbf{if} n % d == 0:
-      \textbf{print}(d)
+      print(d)
 \end{Verbatim}
+
+\vspace{0.5cm}
 		
-			\end{column}
-			\begin{column}	{0.50\textwidth}
-				Answer: \texttt{range(1,n)} generates values up to \texttt{n-1}, so \texttt{n} is not printed. Instead, \texttt{range(1,n+1)} is better. 
-			\end{column}
-			\end{columns}
-			
-				\vspace{2cm}
-				Teachers can often spot such erroneous patterns without test cases. Can we represent their knowledge and use it to:
-				\begin{itemize}
-					\item generate automatic feedback in tutoring systems or
-					\item find programs with similar approach / error?
-				\end{itemize}				
-				Two research questions arise: 
-				\begin{itemize}
-					\item RQ1: How can we encode patterns in programs? 
-					\item RQ2: How can we automatically extract relevant patterns from student solutions?
-				\end{itemize}
-\ No newline at end of file
+Answer: \texttt{range(1,n)} only generates values up to \texttt{n-1}, so \texttt{n} is not printed.
+Instead, \texttt{range(1,n+1)} should be used.
+
+\vspace{1cm}
+
+An experienced teacher can spot such errors easily, without having to run the program. Can we discover that knowledge automatically?
+More specifically:
+
+\vspace{1cm}
+
+\begin{itemize}
+\item How do we encode distinctive program features?
+\item Can we discern which features are relevant to some error or solution?
+\end{itemize}
+
+\vspace{1cm}
+
+We address these questions by using abstract-syntax-tree (AST) patterns to encode program features, and machine-learned rules to discover common errors and solution strategies for beginner Python programs.