Multilevel and Sequential Monte Carlo for Training-Free Diffusion Guidance
A sequential Monte Carlo framework with multilevel variance reduction provides unbiased diffusion guidance, achieving state-of-the-art training-free conditional generation with lower cost.
Published 2026Paris Poster Session 1 · Wed, Dec 9, 12:30 PM–2:30 PM local time · Paris Poster HallarXiv ↗OpenReview ↗
Only vote on papers you've read. Sign in with GitHub to vote.
Abstract
We address the problem of accurate, training-free guidance for conditional generation in trained diffusion models. Existing methods typically rely on point-estimates to approximate the posterior score, often resulting in biased approximations that fail to capture multimodality inherent to the reverse process of diffusion models. We propose a sequential Monte Carlo (SMC) framework that constructs an unbiased estimator of $p_θ(y|x_t)$ by integrating over the full denoising distribution via Monte Carlo approximation. To ensure computational tractability, we incorporate variance-reduction schemes based on Multi-Level Monte Carlo (MLMC). Our approach achieves new state-of-the-art results for training-free guidance on CIFAR-10 class-conditional generation, achieving $95.6\%$ accuracy with $3\times$ lower cost-per-success than baselines. On ImageNet, our algorithm achieves $1.5\times$ cost-per-success advantage over existing methods.