aiqbench analysis
This research introduces PiX-MC, a novel framework that accelerates Bayesian imaging by combining proximal Langevin dynamics with Picard iteration to enable efficient multi-GPU parallelization. By leveraging score-based generative priors within this time-parallel structure, the method achieves up to 50x speedups on large-scale CT reconstruction tasks while maintaining rigorous convergence guarantees for complex posterior distributions. The work represents a significant advancement in scalable inference techniques essential for high-dimensional imaging applications driven by modern AI models.
Picard Proximal Monte Carlo for Parallel Bayesian Imaging with Score-Based Generative Priors
arXiv:2608.17666v1 Announce Type: new Abstract: Bayesian imaging inverse problems often require sampling from high-dimensional posterior distributions. While recent score-based and diffusion models provide expressive Bayesian priors, their sampling procedures remain inherently sequential and computationally expensive for large-scale imaging applications. We propose PiX-MC, a time-parallel posterior sampling framework based on proximal Langevin dynamics and Picard iteration. The proximal-likelihood formulation exploits the fact that many imaging likelihoods admit efficient, problem-specific proximal operators, while Picard refinement exposes parallelism across discretization nodes and naturally supports multi-GPU implementation. To further improve practical scalability and sampling performance, we develop multi-block and annealed variants of the proposed framework. We establish convergence guarantees under transparent assumptions, accommodating non-log-concave posteriors, imperfect learned score models, multi-block implementations, and annealing schedules. Experiments on a diverse collection of imaging inverse problems demonstrate that PiX-MC substantially reduces wall-clock time while preserving reconstruction quality. On a $512\times512\times80$ sparse-view computed tomography (CT) problem, annealed multi-block PiX-MC achieves up to a $50\times$ runtime speedup over the standard Langevin sampler using eight GPUs.
Abstract:Bayesian imaging inverse problems often require sampling from high-dimensional posterior distributions. While recent score-based and diffusion models provide expressive Bayesian priors, their sampling procedures remain inherently sequential and computationally expensive for large-scale imaging applications. We propose PiX-MC, a time-parallel posterior sampling framework based on proximal Langevin dynamics and Picard iteration. The proximal-likelihood formulation exploits the fact that many imaging likelihoods admit efficient, problem-specific proximal operators, while Picard refinement exposes parallelism across discretization nodes and naturally supports multi-GPU implementation. To further improve practical scalability and sampling performance, we develop multi-block and annealed variants of the proposed framework. We establish convergence guarantees under transparent assumptions, accommodating non-log-concave posteriors, imperfect learned score models, multi-block implementations, and annealing schedules. Experiments on a diverse collection of imaging inverse problems demonstrate that PiX-MC substantially reduces wall-clock time while preserving reconstruction quality. On a $512\times512\times80$ sparse-view computed tomography (CT) problem, annealed multi-block PiX-MC achieves up to a $50\times$ runtime speedup over the standard Langevin sampler using eight GPUs.
Machine Learning (cs.LG)
arXiv:2608.17666 [cs.LG]
(or arXiv:2608.17666v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.17666 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 255 words
Related coverage
Posterior sampling by source-space MCMC via prior-based few-step transport maps
arXiv · Oct 2, 2026
BayesPO: Bayesian Prompt Optimization via Parallel-Tempered Gradient-Guided Discrete MCMC
arXiv · Jul 20, 2026
Theoretical and Practical Advances in Hierarchical Bayesian Deep Learning Methodology
arXiv · Oct 15, 2026
Product Monte Carlo: Expectation Under a Product of Densities
arXiv · Oct 12, 2026