Skip to yearly menu bar Skip to main content


Poster

Projected Stein Variational Newton: A Fast and Scalable Bayesian Inference Method in High Dimensions

Peng Chen · Keyi Wu · Joshua Chen · Tom O'Leary-Roseberry · Omar Ghattas

East Exhibition Hall B, C #188

Keywords: [ Probabilistic Methods ] [ Variational Inference ] [ Algorithms -> Kernel Methods; Algorithms -> Nonlinear Dimensionality Reduction and Manifold Learning; Probabilistic Methods ]


Abstract:

We propose a projected Stein variational Newton (pSVN) method for high-dimensional Bayesian inference. To address the curse of dimensionality, we exploit the intrinsic low-dimensional geometric structure of the posterior distribution in the high-dimensional parameter space via its Hessian (of the log posterior) operator and perform a parallel update of the parameter samples projected into a low-dimensional subspace by an SVN method. The subspace is adaptively constructed using the eigenvectors of the averaged Hessian at the current samples. We demonstrate fast convergence of the proposed method, complexity independent of the parameter and sample dimensions, and parallel scalability.

Live content is unavailable. Log in and register to view live content