Regression-based Monte Carlo integration

Abstract

Monte Carlo integration is typically interpreted as an estimator of the expected value using stochastic samples. There exists an alternative interpretation in calculus where Monte Carlo integration can be seen as estimating a constant function---from the stochastic evaluations of the integrand---that integrates to the original integral. The integral mean value theorem states that this constant function should be the mean (or expectation) of the integrand. Since both interpretations result in the same estimator, little attention has been devoted to the calculus-oriented interpretation. We show that the calculus-oriented interpretation actually implies the possibility of using a more complex function than a constant one to construct a more efficient estimator for Monte Carlo integration. We build a new estimator based on this interpretation and relate our estimator to control variates with least-squares regression on the stochastic samples of the integrand. Unlike prior work, our resulting estimator is provably better than or equal to the conventional Monte Carlo estimator. To demonstrate the strength of our approach, we introduce a practical estimator that can act as a simple drop-in replacement for conventional Monte Carlo integration. We experimentally validate our framework on various light transport integrals. The code is available at https://github.com/iribis/regressionmc.

Recursive Control Variates for Inverse Rendering

“…Certain classical algorithms can also be viewed through the lens of control variates, such as residual tracking [Novák et al 2014], delta tracking , multiple importance sampling [Kondapaneni et al 2019], Monte Carlo integration [Salaün et al 2022], and even temporal gradient-domain rendering [Manzi et al 2016], which is equivalent to a particular (unweighted) configuration of the image space control variates of Rousselle et al [2016].…”

An Efficient Self-Tuning Proposal Distribution for Random Variate Generation With Complex Density Representation

“…The associate editor coordinating the review of this manuscript and approving it for publication was Emanuele Crisostomi . frameworks such as entropy estimation [8], reinforcement learning [9], Monte Carlo integration [10], and sample average approximation [11], which expects to use randomly generated discrete scenarios to evaluate a function containing the random variables with complex distribution model, so as to convert the complex uncertainty problems (e.g., stochastic programming [12]) or the large-scale decision processes (e.g., Monte Carlo tree search [13]) into small-scale and easy to handle deterministic problems. In essence, the non-uniform random variate generation provides a solution to simplify the complex nonlinear uncertainty problems involving random variables with certain distributions, to avoid the calculation of the difficult probability model directly, and uses the VOLUME 11, 2023 This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.…”