Skewness of a randomized quasi-Monte Carlo estimate

Zexin Pan,Art B. Owen

2024-05-10

Abstract:Some recent work on confidence intervals for randomized quasi-Monte Carlo (RQMC) sampling found a surprising result: ordinary Student $t$ 95\% confidence intervals based on a modest number of replicates were seen to be very effective and even more reliable than some bootstrap $t$ intervals that were expected to be best. One potential explanation is that those RQMC estimates have small skewness. In this paper we give conditions under which the skewness is $O(n^\epsilon)$ for any $\epsilon>0$, so `almost $O(1)$'. Under a random generator matrix model, we can improve this rate to $O(n^{-1/2+\epsilon})$ with very high probability. We also improve some probabilistic bounds on the distribution of the quality parameter $t$ for a digital net in a prime base under random sampling of generator matrices.

Numerical Analysis,Computation

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: in randomized quasi - Monte Carlo (RQMC) sampling, the influence of the skewness of estimators on the performance of confidence intervals. Specifically, the author explores why the ordinary Student t 95% confidence interval based on a small number of repetitions is very effective in this case, and is even more reliable than the expected better Bootstrap t confidence interval. ### Main problems and background 1. **Validity of confidence intervals**: - The paper points out that in some cases, the ordinary Student t confidence interval performs excellently when based on RQMC estimation, while the Bootstrap t method does not show the expected advantages. - A potential explanation for this phenomenon is that the skewness of the RQMC estimator is small. 2. **Study of skewness**: - The author studies the skewness of the RQMC estimator and proves that under certain conditions, the skewness can reach the order of $ O(n^\epsilon) $ for any $ \epsilon > 0 $, that is, "almost $ O(1) $". - Under the randomly generated matrix model, the skewness can be further improved to $ O(n^{-1/2+\epsilon}) $ and holds with high probability. 3. **Quality parameters of digital nets**: - The author also improves some probability bounds on the quality parameters of digital nets under prime bases, especially in the case of randomly generated matrices. ### Formula representation - Definition of skewness $ \gamma $: \[ \gamma = \frac{E(\hat{\mu}^3)}{E(\hat{\mu}^2)^{3/2}} \] - Definition of kurtosis $ \theta $: \[ \theta = \frac{E(\hat{\mu}^4)}{E(\hat{\mu}^2)^2}- 3 \] ### Conclusion By analyzing the skewness of the RQMC estimator, the author explains why the ordinary Student t confidence interval performs better in this case. In addition, they also provide some theoretical results, indicating that under appropriate conditions, the skewness of the RQMC estimator can be very small, thus improving the reliability of the confidence interval.

Skewness of a randomized quasi-Monte Carlo estimate

Randomized Benchmarking Using Unitary t-Design for Average Fidelity Estimation of Practical Quantum Circuit

Sufficient Conditions for Central Limit Theorems and Confidence Intervals for Randomized Quasi-Monte Carlo Methods

Convergence analysis of quasi-Monte Carlo sampling for quantile and expected shortfall

Achieving High Convergence Rates by Quasi-Monte Carlo and Importance Sampling for Unbounded Integrands

Higher order scrambled digital nets achieve the optimal rate of the root mean square error for smooth integrands

Estimating means of bounded random variables by betting

Guaranteed Conservative Fixed Width Confidence Intervals Via Monte Carlo Sampling

Quasi-Monte Carlo Beyond Hardy-Krause

High Confidence Level Inference is Almost Free using Parallel Stochastic Optimization

ON THE ERROR RATE OF IMPORTANCE SAMPLING WITH RANDOMIZED QUASI-MONTE CARLO*

Randomized quasi-Monte Carlo methods for risk-averse stochastic optimization

Sensitivity estimation of conditional value at risk using randomized quasi-Monte Carlo

Automatic optimal-rate convergence of randomized nets using median-of-means

On the Convergence Rate of Quasi Monte Carlo Method with Importance Sampling for Unbounded Functions in RKHS

On the convergence conditions of Laplace importance sampling with randomized quasi-Monte Carlo

Super-polynomial accuracy of multidimensional randomized nets using the median-of-means

Uncertainty Quantification and Confidence Intervals for Naive Rare-Event Estimators

Convergence of Randomized Quasi-Monte Carlo Sampling for Value-at-Risk and Conditional Value-at-Risk

Randomized and Exchangeable Improvements of Markov's, Chebyshev's and Chernoff's Inequalities

The importance of being scrambled: supercharged Quasi Monte Carlo