\relax 
\newlabel{eq3}{{1}{1}}
\newlabel{eq1}{{2}{1}}
\@writefile{toc}{\contentsline {paragraph}{1. Compute the utility of the geometric mechanism for a counting query, with privacy degree $\epsilon $, on the uniform prior distribution, with the gain function defined as the identity relation}{1}}
\newlabel{eq2}{{3}{1}}
\@writefile{toc}{\contentsline {paragraph}{2. Same exercise, but with the gain function defined as the converse of the distance.}{3}}
\@writefile{toc}{\contentsline {paragraph}{3. Find a mechanism for the same counting query, with the same degree of privacy, but lower utility}{3}}
\@writefile{toc}{\contentsline {paragraph}{4. We saw that post-processing cannot decrease privacy. Can it decrease the utility? Motivate your answer}{4}}