1: \begin{abstract}

2: We study the possibility of using  multilevel algorithms for

3: the computation of correlation functions of gradient flow observables.

4: For each point in the correlation function an approximate flow is defined

5: which depends only on links in a subset of the lattice. Together with 

6: a local action this allows for independent updates and consequently a

7: convergence of the Monte Carlo process faster than the inverse square root of the number

8: of measurements. We demonstrate the feasibility of this idea in the correlation

9: functions of the topological charge and the energy density.

10: \end{abstract}

11: