1: \begin{abstract}%

2: 	In this work we undertake a thorough study of the non-asymptotic properties of the vanilla generative adversarial networks (GANs). We prove a sharp oracle inequality for the Jensen-Shannon (JS) divergence between the underlying density $\pstar$ and the GAN estimate. We also study the rates of convergence in the context of nonparametric density estimation. In particular, we show that the JS-divergence between the GAN estimate and $\pstar$ decays as fast as $(\log{n}/n)^{2\beta/(2\beta+d)}$ where $n$ is the sample size and $\beta$ determines the smoothness of

3: 	$\pstar$. To the best of our knowledge, this is the first result in the literature on density estimation using vanilla

4: 	GANs with JS convergence rates faster than $n^{-1/2}$ in the regime $\beta > d/2$. Moreover, we show that the

5: 	obtained rate is minimax optimal (up to logarithmic factors) for the considered class of densities.

6: \end{abstract}

7: