An Adaptive Importance Sampling for Locally Stable Point Processes

Hee-Geon Kang,Sunggon Kim

2024-08-14

Abstract:The problem of finding the expected value of a statistic of a locally stable point process in a bounded region is addressed. We propose an adaptive importance sampling for solving the problem. In our proposal, we restrict the importance point process to the family of homogeneous Poisson point processes, which enables us to generate quickly independent samples of the importance point process. The optimal intensity of the importance point process is found by applying the cross-entropy minimization method. In the proposed scheme, the expected value of the function and the optimal intensity are iteratively estimated in an adaptive manner. We show that the proposed estimator converges to the target value almost surely, and prove the asymptotic normality of it. We explain how to apply the proposed scheme to the estimation of the intensity of a stationary pairwise interaction point process. The performance of the proposed scheme is compared numerically with the Markov chain Monte Carlo simulation and the perfect sampling.

Machine Learning,Computation

What problem does this paper attempt to address?

The paper attempts to address the problem of finding the expected value of the statistics of a Locally Stable Point Process within a bounded region. Specifically, the paper proposes an adaptive importance sampling method to solve this problem. This method restricts the importance point process to a Homogeneous Poisson Point Process, allowing for the rapid generation of independent sample point processes. Additionally, the paper finds the optimal intensity of the importance point process by applying the cross-entropy minimization method and iteratively estimates the expected value of the function and the optimal intensity in an adaptive manner within the proposed scheme. The paper proves that the proposed estimator almost surely converges to the target value and demonstrates its asymptotic normality. Finally, the paper explores how the scheme can be applied to estimate the intensity of a Stationary Pairwise Interaction Point Process and compares its performance with Markov Chain Monte Carlo Simulation and Perfect Sampling through numerical methods.

An Adaptive Importance Sampling for Locally Stable Point Processes

Efficient large deviation estimation based on importance sampling

Importance sampling for stochastic reaction-diffusion equations in the moderate deviation regime

Local inhomogeneous weighted summary statistics for marked point processes

On importance sampling and independent Metropolis-Hastings with an unbounded weight function

A General Framework for Importance Sampling with Latent Markov Processes

Efficient, Cross-Fitting Estimation of Semiparametric Spatial Point Processes

Risk-neutral limit of adaptive importance sampling of random stopping times

Semi-Parametric Inference for Doubly Stochastic Spatial Point Processes: An Approximate Penalized Poisson Likelihood Approach

Estimation in hidden Markov models via efficient importance sampling

Nonasymptotic Bounds for Suboptimal Importance Sampling

Importance Sampling for the Empirical Measure of Weakly Interacting Diffusions

Efficient Importance Sampling for Rare Event Simulation with Applications

Minimum contrast for the first-order intensity estimation of spatial and spatio-temporal point processes

On a Metropolis-Hastings importance sampling estimator

Implicitly Adaptive Importance Sampling

Sampling Point Processes on Stable Unbounded Regions and Exam Simulation of Queues

Improving control based importance sampling strategies for metastable diffusions via adapted metadynamics

Nonparametric intensity estimation from noisy observations of a Poisson process under unknown error distribution

Optimal sampling design for global approximation of jump diffusion SDEs