Subsampling Error in Stochastic Gradient Langevin Diffusions
This work provides theoretical insights into the error sources of SGLD, which is incremental as it refines existing analyses for subsampling-based MCMC methods in statistical learning.
The paper tackles the problem of analyzing the pure subsampling error in Stochastic Gradient Langevin Dynamics (SGLD) for Bayesian posterior approximation in large-scale data settings, showing that the Wasserstein distance between the posterior and the limiting distribution of an idealized version (SGLDiff) is bounded by a fractional power of the mean waiting time.
The Stochastic Gradient Langevin Dynamics (SGLD) are popularly used to approximate Bayesian posterior distributions in statistical learning procedures with large-scale data. As opposed to many usual Markov chain Monte Carlo (MCMC) algorithms, SGLD is not stationary with respect to the posterior distribution; two sources of error appear: The first error is introduced by an Euler--Maruyama discretisation of a Langevin diffusion process, the second error comes from the data subsampling that enables its use in large-scale data settings. In this work, we consider an idealised version of SGLD to analyse the method's pure subsampling error that we then see as a best-case error for diffusion-based subsampling MCMC methods. Indeed, we introduce and study the Stochastic Gradient Langevin Diffusion (SGLDiff), a continuous-time Markov process that follows the Langevin diffusion corresponding to a data subset and switches this data subset after exponential waiting times. There, we show the exponential ergodicity of SLGDiff and that the Wasserstein distance between the posterior and the limiting distribution of SGLDiff is bounded above by a fractional power of the mean waiting time. We bring our results into context with other analyses of SGLD.