1: \begin{abstract}

2:     The Grenander estimator is a well-studied

3:   procedure for univariate nonparametric density estimation. It is

4:   usually defined as the Maximum Likelihood Estimator (MLE) over the class

5:   of all non-increasing densities on the positive real line. It can

6:   also be seen as the MLE over the class of

7:   all scale mixtures of uniform densities. Using the latter viewpoint,

8:   Pavlides and Wellner~\cite{pavlides2012nonparametric} proposed a

9:   multivariate extension of the  

10:   Grenander estimator as the nonparametric MLE over the class of all

11:   multivariate scale mixtures of uniform densities. We prove that this

12:   multivariate estimator achieves the univariate cube root rate of

13:   convergence with only a logarithmic multiplicative factor that

14:   depends on the dimension. The usual curse of dimensionality is 

15:   therefore avoided to some extent for this multivariate

16:   estimator. This result positively resolves a conjecture of

17:   Pavlides and Wellner~\cite{pavlides2012nonparametric} under an

18:   additional lower bound   assumption. Our proof proceeds via a

19:   general accuracy result for the   Hellinger accuracy of MLEs over

20:   convex classes of densities.   

21: \end{abstract}

22: