1: \begin{abstract} 

2: Stochastic gradient Hamiltonian Monte Carlo (SGHMC) is a variant of stochastic gradient with momentum  where a controlled and properly scaled Gaussian noise is added to the stochastic gradients to steer the iterates towards a global minimum. Many works reported its empirical success in practice for solving stochastic non-convex optimization problems, in particular it has been observed to outperform overdamped Langevin Monte Carlo-based methods such as stochastic gradient Langevin dynamics (SGLD) in many applications. Although asymptotic global convergence properties of SGHMC are well known, its finite-time performance is not well-understood. 

3: In this work, we study two variants of SGHMC based on two alternative discretizations %(explicit Euler and the discretization technique of \cite{Cheng}) 

4: of the underdamped Langevin diffusion.  We provide finite-time performance bounds for the global convergence of both SGHMC variants for solving stochastic non-convex optimization problems with explicit constants. Our results lead to non-asymptotic guarantees for both population and empirical risk minimization problems. For a fixed target accuracy level, on a class of non-convex problems, we obtain complexity bounds for SGHMC that can be tighter than those for SGLD. %up to a square root factor. 

5: These results show that acceleration with momentum is possible in the context of global non-convex optimization. % \add{Mert: "up to a square root factor" still true given the new results in theorem 2 and corollary 3? ignore this comment if it is true..}

6: \end{abstract}

7: