Random Feature Stein Discrepancies

by   Jonathan H. Huggins, et al.

Computable Stein discrepancies have been deployed for a variety of applications, including sampler selection in posterior inference, approximate Bayesian inference, and goodness-of-fit testing. Existing convergence-determining Stein discrepancies admit strong theoretical guarantees but suffer from a computational cost that grows quadratically in the sample size. While linear-time Stein discrepancies have been proposed for goodness-of-fit testing, they exhibit avoidable degradations in testing power---even when power is explicitly optimized. To address these shortcomings, we introduce feature Stein discrepancies (ΦSDs), a new family of quality measures that can be cheaply approximated using importance sampling. We show how to construct ΦSDs that provably determine the convergence of a sample to its target and develop high-accuracy approximations---random ΦSDs (RΦSDs)---which are computable in near-linear time. In our experiments with sampler selection for approximate posterior inference and goodness-of-fit testing, RΦSDs typically perform as well or better than quadratic-time KSDs while being orders of magnitude faster to compute.


page 1

page 2

page 3

page 4


Sampling from Network Reliability in Near-linear Time

We give a O(n) time sampler for independent sets of a matroid with n ele...

Measuring Sample Quality with Kernels

Approximate Markov chain Monte Carlo (MCMC) offers the promise of more r...

Compress Then Test: Powerful Kernel Testing in Near-linear Time

Kernel two-sample testing provides a powerful framework for distinguishi...

Automated Scalable Bayesian Inference via Hilbert Coresets

The automation of posterior inference in Bayesian data analysis has enab...

The inverse of Ackermann function is computable in linear time

We propose a detailed proof of the fact that the inverse of Ackermann fu...

Error bounds for some approximate posterior measures in Bayesian inference

In certain applications involving the solution of a Bayesian inverse pro...

Importance Sampling via Local Sensitivity

Given a loss function F:X→R^+ that can be written as the sum of losses o...

Please sign up or login with your details

Forgot password? Click here to reset