Reducing Complexity of Pseudo-Marginal Sequential Monte Carlo via Hierarchical Tempering

Abstract

The Bayesian approach to inverse problems provides a powerful framework for inferring unknown parameters from noisy observations. Among the computationally most challenging statistical inference problems are problems with intractable likelihoods, especially if no unbiased likelihood estimator is available. We focus on a class of such problems, where a biased likelihood estimator is obtained through a stochastic approximation of a forward map (stochastic solver). Obtaining an empirical posterior approximation in applications is very expensive due to the repeated stochastic solver evaluations required by standard sampling methods. To address this computational bottleneck, we develop a novel algorithm based on a density-tempered marginalized Sequential Monte Carlo (SMC) sampler. Our approach (hierarchical SMC) accelerates inference by leveraging a hierarchy of stochastic solvers with increasing accuracy both along the tempering sequence and within mutation stages of the SMC sampler. We establish, under additional assumptions, that the asymptotic speedup is proportional to the number of tempering stages T. We demonstrate computational gains on several examples, including a nonlinear, high-contrast inverse problem arising in photoacoustic medical imaging. On this problem, we demonstrate a speedup by a factor of 14.3, reducing the inference wall-clock time from about 4.6 days to 7.7 hours on an NVIDIA A40 GPU.

Posterior Tempering Explains Variance Inflation in Linear and Generalized Linear Thompson Sampling
Prateek Jaiswal, Debdeep Pati, Anirban Bhattacharya +1
arXiv2 Sep 2026PreprintCC BY 4.0Full textStatisticsarXiv
The Impact of Likelihood Tempering on the Limiting Predictive Moments of Variational Bayesian Linear Neural Networks
Ian Zhang, Thibault Randrianarisoa
arXiv6 Oct 2026PreprintCC BY 4.0Full textStatisticsarXiv
Delta functions, a local coarea formula and marginal distributions
Ranee Brylinski, Jean-Luc Brylinski
arXiv12 Oct 2026PreprintCC BY 4.0Full textProbabilityarXiv
Multi-marginal optimal transport on sub-Riemannian manifolds
Haowei Lin
arXiv24 Oct 2026PreprintarXiv licenceFull textGeometry & topology Optimisation & controlarXiv
Mimicking Volatilities under Marginal Constraints via q_x-Bass Martingales: A Parallel Computing Approach
Antonio Marini
arXiv22 Oct 2026PreprintarXiv licenceFull textProbabilityarXiv

Cite

BibTeX RIS

BibTeX

@misc{martinek2026reducing,
  title = {{Reducing Complexity of Pseudo-Marginal Sequential Monte Carlo via Hierarchical Tempering}},
  author = {Josef Martínek and Julian Hofstadler and Alexander M. G. Cox and Robert Scheichl},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.37082},
  eprint = {2610.37082},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.37082},
}

RIS

TY  - UNPB
TI  - Reducing Complexity of Pseudo-Marginal Sequential Monte Carlo via Hierarchical Tempering
AU  - Josef Martínek
AU  - Julian Hofstadler
AU  - Alexander M. G. Cox
AU  - Robert Scheichl
PY  - 2026
DA  - 2026/10/27
PB  - arXiv
DO  - 10.48550/arxiv.2610.37082
UR  - https://arxiv.org/abs/2610.37082
AB  - The Bayesian approach to inverse problems provides a powerful framework for inferring unknown parameters from noisy observations. Among the computationally most challenging statistical inference problems are problems with intractable likelihoods, especially if no unbiased likelihood estimator is available. We focus on a class of such problems, where a biased likelihood estimator is obtained through a stochastic approximation of a forward map (stochastic solver). Obtaining an empirical posterior approximation in applications is very expensive due to the repeated stochastic solver evaluations required by standard sampling methods. To address this computational bottleneck, we develop a novel algorithm based on a density-tempered marginalized Sequential Monte Carlo (SMC) sampler. Our approach (hierarchical SMC) accelerates inference by leveraging a hierarchy of stochastic solvers with increasing accuracy both along the tempering sequence and within mutation stages of the SMC sampler. We establish, under additional assumptions, that the asymptotic speedup is proportional to the number of tempering stages T. We demonstrate computational gains on several examples, including a nonlinear, high-contrast inverse problem arising in photoacoustic medical imaging. On this problem, we demonstrate a speedup by a factor of 14.3, reducing the inference wall-clock time from about 4.6 days to 7.7 hours on an NVIDIA A40 GPU.
ER  -

Abstract

Related

Cite