Unreasonable effectiveness of Monte Carlo

This is a comment on the article "Probabilistic Integration: A Role in Statistical Computation?" by F.-X. Briol, C. J. Oates, M. Girolami, M. A. Osborne and D. Sejdinovic to appear in Statistical Science. There is a role for statistical computation in numerical integration. However, the competition from incumbent methods looks to be stiffer for this problem than for some of the newer problems being handled by probabilistic numerics. One of the challenges is the unreasonable effectiveness of the central limit theorem. Another is the unreasonable effectiveness of pseudorandom number generators. A third is the common $O(n^3)$ cost of methods based on Gaussian processes. Despite these advantages, the classical methods are weak in places where probabilistic methods could bring an improvement.

• Publication year

  2019

• Venue

  arXiv.org

• Publication date

  2019-01-18

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 1901.06428
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Intuition for an Old Curiosity and an Implication for MCMC

  2021cited by this paper
• Fast automatic Bayesian cubature using lattice sampling

  2018cited by this paper
• Maximum Likelihood Estimation for a Smooth Gaussian Random Field Model

  2017cited by this paper
• Bayesian Probabilistic Numerical Methods

  2017cited by this paper
• Effective Dimension of Some Weighted Pre-Sobolev Spaces with Dominating Mixed Partial Derivatives

  2017cited by this paper
• Probabilistic numerics and uncertainty in computations

  2015cited by this paper
• Multivariate output analysis for Markov chain Monte Carlo

  2015cited by this paper
• Design and Analysis of Computer Experiments

  2014cited by this paper
• Inference from Simulations and Monitoring Convergence

  2011influential reference
• Higher order scrambled digital nets achieve the optimal rate of the root mean square error for smooth integrands

  2010influential reference
• An Introduction to the Bootstrap

  2007cited by this paper
• Finite-order Integration Weights Can be Dangerous

  2007cited by this paper
• TestU01: A C library for empirical testing of random number generators

  2006cited by this paper
• On a Likelihood Approach for Monte Carlo Integration

  2004cited by this paper
• EMPIRICAL-LIKELIHOOD-BASED CONFIDENCE INTERVAL FOR THE MEAN WITH A HEAVY-TAILED DISTRIBUTION

  2004influential reference
• Average-case analysis of numerical problems

  2000cited by this paper
• Bayesian numerical analysis

  2000influential reference
• Design and analysis of computer experiments

  1998cited by this paper
• An Introduction to the Bootstrap

  1995cited by this paper
• Monte Carlo simulations: Hidden errors from "good" random number generators.

  1992cited by this paper
• Random number generation and Quasi-Monte Carlo methods

  1992cited by this paper
• Empirical Likelihood and Small Samples

  1992cited by this paper
• Theoretical Comparison of Bootstrap Confidence Intervals

  1988cited by this paper
• Non-Uniform Random Variate Generation

  1986cited by this paper
• On the consistency of Bayes estimates

  1986cited by this paper
• Quasi-Monte Carlo methods and pseudo-random numbers

  1978cited by this paper
• A Pseudo-Random Number Generator for the System/360

  1969cited by this paper

• A modern retrospective on probabilistic numerics

  2019cites this paper