1: \begin{abstract}

2: Many recent Markov chain Monte Carlo (MCMC) samplers leverage continuous dynamics to define a transition kernel that efficiently explores a target distribution. In tandem, a focus has been on devising scalable variants that subsample the data and use stochastic gradients in place of full-data gradients in the dynamic simulations. However, such stochastic gradient MCMC samplers have lagged behind their full-data counterparts in terms of the complexity of dynamics considered since proving convergence in the presence of the stochastic gradient noise is non-trivial.  Even with simple dynamics, significant physical intuition is often required to modify the dynamical system to account for the stochastic gradient noise.  In this paper, we provide a general recipe for constructing MCMC samplers---including stochastic gradient versions---based on continuous Markov processes specified via two matrices.  We constructively prove that the framework is \emph{complete}. That is, any continuous Markov process that provides samples from the target distribution can be written in our framework.  We show how previous continuous-dynamic samplers can be trivially ``reinvented'' in our framework, avoiding the complicated sampler-specific proofs. We likewise use our recipe to straightforwardly propose a new state-adaptive sampler: \emph{stochastic gradient Riemann Hamiltonian Monte Carlo} (SGRHMC).  Our experiments on simulated data and a streaming Wikipedia analysis demonstrate that the proposed SGRHMC sampler inherits the benefits of Riemann HMC, with the scalability of stochastic gradient methods.

3: \end{abstract}

4: