Tempered Guided Diffusion

TL;DR

Tempered Guided Diffusion introduces an annealed Monte Carlo method for efficient, training-free conditional sampling with diffusion models, improving posterior approximation and computational efficiency.

Contribution

It proposes a novel Tempered Guided Diffusion framework that adaptively concentrates computation on plausible trajectories, enhancing sampling quality without task-specific training.

Findings

TGD achieves better posterior approximation than baseline methods.

A-TGD effectively prunes trajectories, reducing computation while maintaining quality.

Experiments demonstrate improved speed-quality tradeoffs in inverse problems.

Abstract

Training-free conditional diffusion provides a flexible alternative to task-specific conditional model training, but existing samplers often allocate computation inefficiently: independent guided trajectories can vary widely in quality, and additional function evaluations along a single trajectory may not recover from poor early decisions. We propose Tempered Guided Diffusion (TGD), an annealed sequential Monte Carlo framework for training-free conditional sampling with diffusion priors. TGD targets tempered posterior distributions over the clean signal, using noisy diffusion states only as auxiliary variables for proposing reconstructions and propagating particles. Particles are reweighted by incremental likelihood ratios, resampled, and propagated across noise levels, concentrating computation on trajectories plausible under both the prior and observation. Under idealized exact-reconstruction assumptions, full TGD yields a consistent particle approximation to the posterior as the number of particles grows. For expensive reconstruction tasks, Accelerated TGD (A-TGD) retains early particle exploration but prunes to a single high-likelihood trajectory partway through sampling. Experiments on a controlled two-dimensional inverse problem and image inverse problems show improved posterior approximation and favorable wall-clock speed-quality tradeoffs over independent multi-trajectory baselines.