The paper introduces the Energy-based Feynman-Kac Corrector (EBFKC), a method that improves inference-time scaling for diffusion models that are imperfect due to data and training limitations. EBFKC derives Feynman-Kac dynamics to exactly track a prescribed path in continuous time, then approximates these dynamics with sequential Monte Carlo and variance‑controlling guidance. Experiments on Gaussian mixtures, particle systems, and molecular simulations demonstrate that EBFKC closely matches target distributions and free‑energy profiles, outperforming standard inference‑time scaling baselines that still exhibit significant sampling errors.
By Zuokai Wen, Louis Grenioux, Weinan E, Jiequn Han
arXiv:2608. 15144v1 Announce Type: cross Abstract: Posterior sampling with a pretrained diffusion prior is governed by a conditional score whose intermediate likelihood component is generally intractable.
By Zhaoqiang Liu, Tongyao Pang, Ruibing Wang, Yang Zheng
arXiv:2601. 21026v2 Announce Type: replace-cross Abstract: Sampling configurations at thermodynamic equilibrium is a central challenge in statistical physics.
By Louis Grenioux, Maxence Noble
arXiv:2609.00279v1 Announce Type: cross
Abstract: This work shows that diffusion models learned with standard denoising loss can provide effective global MCMC proposals for complex high-dimensional t...
By Mitch Hill
arXiv:2607.04780v2 Announce Type: replace-cross
Abstract: Post-hoc conditioning of pretrained diffusion models can be addressed using Sequential Monte Carlo (SMC) methods. By evolving an interacting...
By Stanislas Strasman (SU, LPSM), Gabriel Victorino Cardoso (STIM), Sylvain Le Corff (LPSM), Vincent Lemaire (LPSM), Antonio Ocello
The paper introduces a method for specificity‑aware diffusion steering that suppresses undesired samples while preserving desired ones. By formulating the problem as a target‑design task, it derives a time‑dependent target distribution based on overlap between positive and negative reference distributions, and samples from it using a variance‑reduced Sequential Monte Carlo (SMC) sampler. Experiments on synthetic, class‑contrastive, text‑to‑image, and peptide‑MHC tasks demonstrate reduced mode shift, improved sampling stability, and better suppression of undesired regions compared to negative‑guidance baselines.
By Luran Wang, Linrui Ma, Hannes St\"ark, Regina Barzilay
We address the problem of efficiently sampling multimodal probability distributions, where standard Markov Chain Monte Carlo methods often suffer from poor mixing and mode trapping. To mitigate these issues, we propose Gradient-free Riemannian Langevin Sampler (GRiLS), a novel proposal that improves exploration without requiring gradient evaluations of the target density.
arXiv:2607. 07519v1 Announce Type: new Abstract: We address the problem of efficiently sampling multimodal probability distributions, where standard Markov Chain Monte Carlo methods often suffer from poor mixing and mode trapping.
By Ricardo Baptista, Olivier Zahm
The paper introduces the concept of local geometric mixing, which focuses on geometric convergence to equilibrium in total variation over a finite number of transitions, thereby capturing rapid local equilibration even when global mixing is slow. It derives local geometric mixing bounds using Dobrushin contraction and applies these results to Diffusion Path Monte Carlo, a Markov chain Monte Carlo method that aligns with the Proximal Sampler. The analysis provides mixing guarantees for both the ideal method and its Metropolis-adjusted implementation, and extends recent spectral gap estimates into mixing time bounds.
By Stefan Oberd\"orster
arXiv:2607. 15682v1 Announce Type: new Abstract: Sampling from an unnormalized Boltzmann density requires proposals that move probability mass globally while retaining enough path-probability information for statistical correction.
By Moxian Qian
arXiv:2603. 27996v2 Announce Type: replace Abstract: Diffusion models have emerged as a powerful framework for generative tasks in deep learning.
By Nihal Sanjay Singh, Mazdak Mohseni-Rajaee, Shaila Niazi, Kerem Y. Camsari
arXiv:2506. 21511v2 Announce Type: replace-cross Abstract: We develop sampling methods, which consist of Gaussian invariant versions of random walk Metropolis (RWM), Metropolis adjusted Langevin algorithm (MALA) and second order Hessian or Manifold MALA.
By Michalis K. Titsias, Angelos Alexopoulos, Siran Liu, Petros Dellaportas