Merge branch 'master' of https://github.com/jeanpouget-abadie/learnNoptimize

1 files changed, 6 insertions, 0 deletions

diff --git a/results.tex b/results.tex
index 1848d35..c9acedd 100644
--- a/results.tex
+++ b/results.tex
@@ -314,6 +314,12 @@ learning $f$ approximately to optimize it.
 
 \subsection{Parametric submodular functions}
 
+\textbf{Note:} all known examples of submodular functions are parametric. The
+question is not parametric vs non-parametric but whether or not the number of
+parameters is polynomial or exponential in $n$ (the size of the ground set).
+For all examples I know except martroid rank functions, it seems to be
+polynomial.
+
 \subsubsection{Influence Functions}
 
 Let $G = (V, E)$ be a weighted directed graph. Without loss of generality, we