Nips format and minor tweaks. We are already close to the page limit…

1 files changed, 1 insertions, 1 deletions

diff --git a/paper/sections/negative.tex b/paper/sections/negative.tex
index f88117a..2eaeb47 100644
--- a/paper/sections/negative.tex
+++ b/paper/sections/negative.tex
@@ -25,7 +25,7 @@ guarantees are sufficient for optimization. First look at learning weights in
 a cover function. Maybe facility location? Sums of concave over modular are
 probably too hard because of the connection to neural networks.
 
-\subsection{Optimizable $n\Rightarrow$ Learnable}
+\subsection{Optimizable $\nRightarrow$ Learnable}
 
 Recall the matroid construction from~\cite{balcan2011learning}:
 \begin{theorem}