1: \begin{abstract}

2: \noindent  

3: In this work we consider  Wasserstein barycenters (average in Wasserstein distance) in Fourier basis. We prove that random Fourier parameters of the barycenter converge to Gaussian random vector by distribution. The convergence approximation has been done in finite-sample condition with  rate $O(p/\sqrt{n})$ depending on measures count ($n$) and the dimension of parameters ($p$). 

4: 

5: Keywords: GAR, Wasserstein distance, multivariate central limit theorem, statistical learning, convex analysis. 

6: 

7: \end{abstract}

8: