1: \begin{abstract}

2: We propose an adaptively weighted stochastic gradient Langevin dynamics algorithm (SGLD), so-called contour stochastic gradient Langevin dynamics (CSGLD), for Bayesian learning in big data statistics. The proposed algorithm is scalable, which automatically adjusts the target distribution such that the simulation and global optimization for a multi-modal distribution can be greatly facilitated. Theoretically, we establish the convergence of the algorithm. The CSGLD algorithm is tested on multiple benchmark datasets including CIFAR10 and CIFAR100. The numerical results indicate its superiority over the existing state-of-the-art algorithms in training deep neural networks.

3: \end{abstract}

4: