# Semi-Exact Control Functionals From Sard’s Method

@article{South2020SemiExactCF, title={Semi-Exact Control Functionals From Sard’s Method}, author={Leah South and Toni Karvonen and Christopher Nemeth and Mark A. Girolami and Chris J. Oates}, journal={arXiv: Computation}, year={2020} }

This paper focuses on the numerical computation of posterior expected quantities of interest, where existing approaches based on ergodic averages are gated by the asymptotic variance of the integrand. To address this challenge, a novel variance reduction technique is proposed, based on Sard's approach to numerical integration and the control functional method. The use of Sard's approach ensures that our control functionals are exact on all polynomials up to a fixed degree in the Bernstein-von… Expand

#### Figures from this paper

#### 9 Citations

Scalable Control Variates for Monte Carlo Methods via Stochastic Optimization

- Mathematics, Computer Science
- ArXiv
- 2020

This paper considers control variates based on Stein operators, presenting a framework that encompasses and generalizes existing approaches that use polynomials, kernels and neural networks, leading to scalable and effective control Variates. Expand

A Riemann–Stein Kernel Method

- 2021

This paper proposes and studies a numerical method for approximation of posterior expectations based on interpolation with a Stein reproducing kernel. Finite-sample-size bounds on the approximation… Expand

Vector-Valued Control Variates

- Mathematics
- 2021

Control variates are post-processing tools for Monte Carlo estimators which can lead to significant variance reduction. This approach usually requires a large number of samples, which can be… Expand

Post-Processing of MCMC

- Computer Science
- 2021

State-of-the-art techniques for post-processing Markov chain output are reviewed, including methods based on discrepancy minimisation, which directly address the bias-variance trade-off, as well as general-purpose control variate methods for approximating expected quantities of interest. Expand

Kernel Minimum Divergence Portfolios

- Computer Science, Mathematics
- ArXiv
- 2021

This paper proves the analytic computability of the underlying mean embedding for various target distribution-kernel pairs, and shows that such analytic knowledge can lead to faster convergence of MMD estimators, and extends the results to the unbounded exponential kernel with minimax lower bounds. Expand

Positively Weighted Kernel Quadrature via Subsampling

- Mathematics, Computer Science
- ArXiv
- 2021

Numerical experiments indicate that the theoretical analysis combines the spectral properties of the kernel with random sampling of points results in effective algorithms to construct kernel quadrature rules with positive weights and small worst-case error. Expand

Minimum Discrepancy Methods in Uncertainty Quantification

- Mathematics
- 2021

5 Partial Solutions to Exercises 29 5.1 Exercise 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 5.2 Exercise 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . .… Expand

Postprocessing of MCMC

- Mathematics
- Annual Review of Statistics and Its Application
- 2021

Markov chain Monte Carlo is the engine of modern Bayesian statistics, being used to approximate the posterior and derived quantities of interest. Despite this, the issue of how the output from a… Expand

Optimal Thinning of MCMC Output

- Computer Science, Mathematics
- 2020

A novel method is proposed, based on greedy minimisation of a kernel Stein discrepancy, that is suitable for problems where heavy compression is required and its effectiveness is demonstrated in the challenging context of parameter inference for ordinary differential equations. Expand

#### References

SHOWING 1-10 OF 84 REFERENCES

A Bayes-Sard Cubature Method

- Computer Science, Mathematics
- NeurIPS
- 2018

Bayes-Sard cubature is introduced, a probabilistic framework that combines the flexibility of Bayesian cubature with the robustness of classical cubatures which are well-established, and two orders of magnitude reduction in error are reported in the context of a high-dimensional financial integral. Expand

Control functionals for Monte Carlo integration

- Mathematics
- 2014

Summary
A non-parametric extension of control variates is presented. These leverage gradient information on the sampling density to achieve substantial variance reduction. It is not required that… Expand

Convergence rates for a class of estimators based on Stein’s method

- Mathematics
- Bernoulli
- 2019

Gradient information on the sampling distribution can be used to reduce the variance of Monte Carlo estimators via Stein's method. An important application is that of estimating an expectation of a… Expand

Scalable Control Variates for Monte Carlo Methods via Stochastic Optimization

- Mathematics, Computer Science
- ArXiv
- 2020

This paper considers control variates based on Stein operators, presenting a framework that encompasses and generalizes existing approaches that use polynomials, kernels and neural networks, leading to scalable and effective control Variates. Expand

Variance reduction for MCMC methods via martingale representations

- Mathematics
- 2019

In this paper we propose an efficient variance reduction approach for additive functionals of Markov chains relying on a novel discrete time martingale representation. Our approach is fully… Expand

A Riemann–Stein Kernel Method

- 2021

This paper proposes and studies a numerical method for approximation of posterior expectations based on interpolation with a Stein reproducing kernel. Finite-sample-size bounds on the approximation… Expand

Regularised Zero-Variance Control Variates for High-Dimensional Variance Reduction

- Mathematics
- 2018

Zero-variance control variates (ZV-CV) are a post-processing method to reduce the variance of Monte Carlo estimators of expectations using the derivatives of the log target. Once the derivatives are… Expand

Bayes–Hermite quadrature

- Mathematics
- 1991

Abstract Bayesian quadrature treats the problem of numerical integration as one of statistical inference. A prior Gaussian process distribution is assumed for the integrand, observations arise from… Expand

The reproducing Stein kernel approach for post-hoc corrected sampling

- Mathematics
- 2020

Stein importance sampling is a widely applicable technique based on kernelized Stein discrepancy, which corrects the output of approximate sampling algorithms by reweighting the empirical… Expand

Diffusion approximations and control variates for MCMC

- Mathematics
- 2018

A new methodology is presented for the construction of control variates to reduce the variance of additive functionals of Markov Chain Monte Carlo (MCMC) samplers. Our control variates are defined as… Expand