Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2504.18677 · Information and Inference A Journal of the IMA

Empirical Bernstein and betting confidence intervals for randomized quasi-Monte Carlo

Aadit Jain, Fred J. Hickernell, Art B. Owen, Aleksei G. Sorokin

Published Feb 6, 2026• 3 citations
arXivHTMLPDF

Abstract

Abstract Randomized quasi-Monte Carlo (RQMC) methods estimate the mean of a random variable by sampling an integrand at $n$ equidistributed points. For scrambled digital nets, the resulting variance is typically $\tilde O(n^{-\theta })$, where $\theta \in [1,3]$ depends on the smoothness of the integrand and $\tilde O$ neglects logarithmic factors. While RQMC can be far more accurate than plain Monte Carlo (MC), it remains difficult to get confidence intervals on RQMC estimates. We investigate some empirical Bernstein confidence intervals (EBCIs) and hedged betting confidence intervals (HBCIs), both from Waudby-Smith and Ramdas (2024, J. Roy. Statist. Soc. B, 86, 1–27), when the random variable of interest is subject to known bounds. When there are $N$ integrand evaluations partitioned into $R$ independent replicates of $n=N/R$ RQMC points, and the RQMC variance is $\varTheta (n^{-\theta })$, then an oracle minimizing the width of a Bennett confidence interval would choose $n =\varTheta (N^{1/(\theta +1)})$. The resulting intervals have a width $\varTheta (N^{-\theta /(\theta +1)})$. Our empirical investigations had optimal values of $n$ grow slowly with $N$, HBCI intervals that were usually narrower than the EBCI ones and optimal values of $n$ for HBCI that were equal to or smaller than the ones for the oracle.

DOI
10.1093/imaiai/iaag003

Related papers

On the error distribution for randomly-shifted lattice rules

Pierre L’Ecuyer, Bruno Tuffin

Randomized quasi-Monte Carlo (RQMC) methods estimate the expectation of a random variable by the average of n dependent realizations of it. In general, due to the strong dependence, the estimation error may not obey a central limit theorem. Analysis of RQMC methods have so far focused mostly on the convergence rates of asymptotic worst-case error bounds and variance bounds, when n ¿ ¿, but little is known about the limiting distribution of the error. Here we examine this limiting distribution for the special case of a randomly-shifted lattice rule, when the integrand is smooth. We start with simple one-dimensional functions, where we show that the limiting distribution is uniform over a bounded interval if the integrand is non-periodic, and has a square root form over a bounded interval if the integrand is periodic. In higher dimensions, for linear functions, the distribution function of the properly standardized error converges to a spline of degree equal to the dimension.

Winter Simulation Conference•Dec 1, 2009•2 citations

Some convergence results in randomized quasi Monte Carlo

Yang Liu

We study the convergence rate of randomized quasi-Monte Carlo (RQMC) methods for integrands with different smoothness. The Koksma--Hlawka inequality fails to be informative for certain $L^2$ singular integrands. We provide convergence results based on spectral analysis.

•Jan 1, 2025•0 citations

Quasi-Monte Carlo Beyond Hardy-Krause II: $(1 + \varepsilon)n$ Samples Suffice

Ekene Ezeunala, Agastya Vibhuti Jha, Haotian Jiang

Numerical integration studies how well one can estimate the integral of a function $f$ over $[0,1)^d$ using $n$ sample points. The two classical methods, Monte Carlo (MC) and quasi-Monte Carlo (QMC), have complementary strengths and weaknesses, and a fundamental question is to design an approach that combines the benefits of both. Recently, building on the transference principle in discrepancy theory, Bansal and Jiang~\cite{BJ25a} gave a randomized QMC method that bridges MC and QMC guarantees using only i.i.d.\ samples. Their method also goes beyond the classical Koksma--Hlawka inequality: it achieves integration error $\widetilde{O}_d(σ_{\mathsf{SO}}(f)/n)$, where the smoothed-out variation $σ_{\mathsf{SO}}(f)$ can be substantially smaller than the Hardy--Krause variation that governs the classical bound. However, their algorithm requires $n^2$ i.i.d.\ samples as input, and this quadratic blowup is inherent to any method based on the transference principle. In this work, we bypass the quadratic blowup: for any constant $\varepsilon > 0$, we show that $(1+\varepsilon)n$ i.i.d.\ samples suffice to both obtain the beyond-Hardy--Krause guarantee of~\cite{BJ25a}, resolving an open problem posed there, and to produce low-discrepancy point sequences. Our algorithms are variants of the online Haar-thinning method of Dwivedi, Feldheim, Gurel-Gurevich, and Ramdas~\cite{DFG+19}.

arXiv:2609.10921•Sep 10, 2026•0 citations

On the Convergence Rate of Randomized Quasi--Monte Carlo for Discontinuous Functions

Zhijian He, Xiaoqun Wang

This paper studies the convergence rate of randomized quasi--Monte Carlo (RQMC) for discontinuous functions, which are often of infinite variation in the sense of Hardy and Krause. It was previously known that the root mean square error (RMSE) of RQMC is only $o(n^{-1/2})$ for discontinuous functions. For certain discontinuous functions in $d$ dimensions, we prove that the RMSE of RQMC is $O(n^{-1/2-1/(4d-2)+\epsilon})$ for any $\epsilon>0$ and arbitrary $n$. If some discontinuity boundaries are parallel to some coordinate axes, the rate can be improved to $O(n^{-1/2-1/(4d_u-2)+\epsilon})$, where $d_u$ denotes the so-called irregular dimension, that is, the number of axes which are not parallel to the discontinuity boundaries. Moreover, this paper shows that the RMSE is $O(n^{-1/2-1/(2d)})$ for certain indicator functions.

SIAM Journal on Numerical Analysis•Jan 1, 2015•28 citations

Quasi-Monte Carlo integration over non-cubical domains

Nilofar Mortazavi

Monte Carlo (MQ) method is a powerful tool to approximate high dimensional integrals. The disadvantage of ordinary MQ is the slow convergence rate by cause of the essential randomness of this method. Computations of this convergence can lead to pure time consuming. Quasi-Monte Carlo (QMC) method yields considerably better results since this method is a deterministic alternative which uses Low-Discrepancy sequences instead of random samples . There are a multiplicity of big open problems in QMC-methods, problems partly arising from applications and partly arising from theory. QMC are developed to integrate over unite cube, where it has much more accuracy than MC for integrands of bounded variation. Integration over more general spaces such as triangles, disks and Cartesian products of such spaces is more challenging for QMC. Nevertheless in real-world applications various problems are defined over such spaces. The aim of this thesis is to provide a survey of a solution of such problems of numerical integration defined over non-cubical spaces. We present QMC and randomised QMC (RQMC) constructions in the triangle with a vanishing discrepancy based on the recent work of Basu 2016. The QMC construction is a version of the Van der Corput-Halton sequences specially made to the unit triangle. The attraction of scrambled net is replication based error estimation for QMC with slightly the same accuracy as QMC, and for smooth enough integrands.

•Jan 1, 2017•0 citations

Robust high-dimensional integration using medians of coarsely scrambled Sobol' sequences

Ziyang Ye, Chaokun Zhu, Zexin Pan

We study the numerical approximation of high-dimensional integrals using randomized quasi-Monte Carlo (RQMC) methods, with a focus on scrambled Sobol' sequences. While asymptotically faster than Monte Carlo, classical RQMC suffers from error bounds that grow exponentially in the dimension $s$, and its root mean squared error (RMSE) convergence rate is generally no better than $O(N^{-3/2})$. To overcome these limitations, we combine two recent developments: Suzuki's coarse scrambling and the median trick. Coarse scrambling randomizes Sobol' sequences according to their generating base polynomials and significantly reduces the maximal gain coefficient. By appropriately choosing the base polynomials, we show that the coefficient can be made uniformly bounded in $s$, yielding an $O(N^{-1/2})$ RMSE for $L^2$ integrands with a dimension-independent constant. The median trick then enables near-optimal convergence for function classes beyond $L^2$: for $L^p$ integrands with $p\in(1,2)$, the median of independent coarsely scrambled estimates achieves an error of $O(N^{-1+1/p})$ with high probability; for integrands in the Haar wavelet space $\mathcal H_{\mathrm{wav},α,s,p,q}$ with $α_p:=α-(1/p-1/2)_+>0$, we prove a high-probability error bound of $O(N^{-α_p-1/2+\varepsilon})$ for any $\varepsilon>0$; the bound improves to $O(N^{-r-α_p-1/2+\varepsilon})$ when the integrand has dominating mixed derivatives of order $r\ge1$ that belong to $\mathcal H_{\mathrm{wav},α,s,p,q}$. The latter two bounds are uniform in $s$ under suitable conditions on the ANOVA components of the integrands. Numerical experiments confirm the predicted convergence rates and demonstrate the robustness of the proposed approach.

arXiv:2609.05979•Sep 5, 2026•0 citations
Abstract Randomized quasi-Monte Carlo (RQMC) methods estimate the mean of a random variable by sampling an integrand at nnn equidistributed points. For scrambled digital nets, the resulting variance is typically O~(n−θ)\tilde O(n^{-\theta })O~(n−θ), where θ∈[1,3]\theta \in [1,3]θ∈[1,3] depends on the smoothness of the integrand and O~\tilde OO~ neglects logarithmic factors. While RQMC can be far more accurate than plain Monte Carlo (MC), it remains difficult to get confidence intervals on RQMC estimates. We investigate some empirical Bernstein confidence intervals (EBCIs) and hedged betting confidence intervals (HBCIs), both from Waudby-Smith and Ramdas (2024, J. Roy. Statist. Soc. B, 86, 1–27), when the random variable of interest is subject to known bounds. When there are NNN integrand evaluations partitioned into RRR independent replicates of n=N/Rn=N/Rn=N/R RQMC points, and the RQMC variance is Θ(n−θ)\varTheta (n^{-\theta })Θ(n−θ), then an oracle minimizing the width of a Bennett confidence interval would choose n=Θ(N1/(θ+1))n =\varTheta (N^{1/(\theta +1)})n=Θ(N1/(θ+1)). The resulting intervals have a width Θ(N−θ/(θ+1))\varTheta (N^{-\theta /(\theta +1)})Θ(N−θ/(θ+1)). Our empirical investigations had optimal values of nnn grow slowly with NNN, HBCI intervals that were usually narrower than the EBCI ones and optimal values of nnn for HBCI that were equal to or smaller than the ones for the oracle.
Randomized quasi-Monte Carlo (RQMC) methods estimate the expectation of a random variable by the average of n dependent realizations of it. In general, due to the strong dependence, the estimation error may not obey a central limit theorem. Analysis of RQMC methods have so far focused mostly on the convergence rates of asymptotic worst-case error bounds and variance bounds, when n ¿ ¿, but little is known about the limiting distribution of the error. Here we examine this limiting distribution for the special case of a randomly-shifted lattice rule, when the integrand is smooth. We start with simple one-dimensional functions, where we show that the limiting distribution is uniform over a bounded interval if the integrand is non-periodic, and has a square root form over a bounded interval if the integrand is periodic. In higher dimensions, for linear functions, the distribution function of the properly standardized error converges to a spline of degree equal to the dimension.
We study the convergence rate of randomized quasi-Monte Carlo (RQMC) methods for integrands with different smoothness. The Koksma--Hlawka inequality fails to be informative for certain L2L^2L2 singular integrands. We provide convergence results based on spectral analysis.
Quasi-Monte Carlo Beyond Hardy-Krause II: (1+ε)n(1 + \varepsilon)n(1+ε)n Samples Suffice
Numerical integration studies how well one can estimate the integral of a function fff over [0,1)d[0,1)^d[0,1)d using nnn sample points. The two classical methods, Monte Carlo (MC) and quasi-Monte Carlo (QMC), have complementary strengths and weaknesses, and a fundamental question is to design an approach that combines the benefits of both. Recently, building on the transference principle in discrepancy theory, Bansal and Jiang~\cite{BJ25a} gave a randomized QMC method that bridges MC and QMC guarantees using only i.i.d.\ samples. Their method also goes beyond the classical Koksma--Hlawka inequality: it achieves integration error O~d(σSO(f)/n)\widetilde{O}_d(σ_{\mathsf{SO}}(f)/n)Od​(σSO​(f)/n), where the smoothed-out variation σSO(f)σ_{\mathsf{SO}}(f)σSO​(f) can be substantially smaller than the Hardy--Krause variation that governs the classical bound. However, their algorithm requires n2n^2n2 i.i.d.\ samples as input, and this quadratic blowup is inherent to any method based on the transference principle. In this work, we bypass the quadratic blowup: for any constant ε>0\varepsilon > 0ε>0, we show that (1+ε)n(1+\varepsilon)n(1+ε)n i.i.d.\ samples suffice to both obtain the beyond-Hardy--Krause guarantee of~\cite{BJ25a}, resolving an open problem posed there, and to produce low-discrepancy point sequences. Our algorithms are variants of the online Haar-thinning method of Dwivedi, Feldheim, Gurel-Gurevich, and Ramdas~\cite{DFG+19}.
This paper studies the convergence rate of randomized quasi--Monte Carlo (RQMC) for discontinuous functions, which are often of infinite variation in the sense of Hardy and Krause. It was previously known that the root mean square error (RMSE) of RQMC is only o(n−1/2)o(n^{-1/2})o(n−1/2) for discontinuous functions. For certain discontinuous functions in ddd dimensions, we prove that the RMSE of RQMC is O(n−1/2−1/(4d−2)+ϵ)O(n^{-1/2-1/(4d-2)+\epsilon})O(n−1/2−1/(4d−2)+ϵ) for any ϵ>0\epsilon>0ϵ>0 and arbitrary nnn. If some discontinuity boundaries are parallel to some coordinate axes, the rate can be improved to O(n−1/2−1/(4du−2)+ϵ)O(n^{-1/2-1/(4d_u-2)+\epsilon})O(n−1/2−1/(4du​−2)+ϵ), where dud_udu​ denotes the so-called irregular dimension, that is, the number of axes which are not parallel to the discontinuity boundaries. Moreover, this paper shows that the RMSE is O(n−1/2−1/(2d))O(n^{-1/2-1/(2d)})O(n−1/2−1/(2d)) for certain indicator functions.
Monte Carlo (MQ) method is a powerful tool to approximate high dimensional integrals. The disadvantage of ordinary MQ is the slow convergence rate by cause of the essential randomness of this method. Computations of this convergence can lead to pure time consuming. Quasi-Monte Carlo (QMC) method yields considerably better results since this method is a deterministic alternative which uses Low-Discrepancy sequences instead of random samples . There are a multiplicity of big open problems in QMC-methods, problems partly arising from applications and partly arising from theory. QMC are developed to integrate over unite cube, where it has much more accuracy than MC for integrands of bounded variation. Integration over more general spaces such as triangles, disks and Cartesian products of such spaces is more challenging for QMC. Nevertheless in real-world applications various problems are defined over such spaces. The aim of this thesis is to provide a survey of a solution of such problems of numerical integration defined over non-cubical spaces. We present QMC and randomised QMC (RQMC) constructions in the triangle with a vanishing discrepancy based on the recent work of Basu 2016. The QMC construction is a version of the Van der Corput-Halton sequences specially made to the unit triangle. The attraction of scrambled net is replication based error estimation for QMC with slightly the same accuracy as QMC, and for smooth enough integrands.
We study the numerical approximation of high-dimensional integrals using randomized quasi-Monte Carlo (RQMC) methods, with a focus on scrambled Sobol' sequences. While asymptotically faster than Monte Carlo, classical RQMC suffers from error bounds that grow exponentially in the dimension sss, and its root mean squared error (RMSE) convergence rate is generally no better than O(N−3/2)O(N^{-3/2})O(N−3/2). To overcome these limitations, we combine two recent developments: Suzuki's coarse scrambling and the median trick. Coarse scrambling randomizes Sobol' sequences according to their generating base polynomials and significantly reduces the maximal gain coefficient. By appropriately choosing the base polynomials, we show that the coefficient can be made uniformly bounded in sss, yielding an O(N−1/2)O(N^{-1/2})O(N−1/2) RMSE for L2L^2L2 integrands with a dimension-independent constant. The median trick then enables near-optimal convergence for function classes beyond L2L^2L2: for LpL^pLp integrands with p∈(1,2)p\in(1,2)p∈(1,2), the median of independent coarsely scrambled estimates achieves an error of O(N−1+1/p)O(N^{-1+1/p})O(N−1+1/p) with high probability; for integrands in the Haar wavelet space Hwav,α,s,p,q\mathcal H_{\mathrm{wav},α,s,p,q}Hwav,α,s,p,q​ with αp:=α−(1/p−1/2)+>0α_p:=α-(1/p-1/2)_+>0αp​:=α−(1/p−1/2)+​>0, we prove a high-probability error bound of O(N−αp−1/2+ε)O(N^{-α_p-1/2+\varepsilon})O(N−αp​−1/2+ε) for any ε>0\varepsilon>0ε>0; the bound improves to O(N−r−αp−1/2+ε)O(N^{-r-α_p-1/2+\varepsilon})O(N−r−αp​−1/2+ε) when the integrand has dominating mixed derivatives of order r≥1r\ge1r≥1 that belong to Hwav,α,s,p,q\mathcal H_{\mathrm{wav},α,s,p,q}Hwav,α,s,p,q​. The latter two bounds are uniform in sss under suitable conditions on the ANOVA components of the integrands. Numerical experiments confirm the predicted convergence rates and demonstrate the robustness of the proposed approach.