Back to Explorer
Research PaperResearchia:202601.100e8400[Data Science > Data Science]

Inference-Time Alignment for Diffusion Models via Doob's Matching

Jinyuan Chang

Abstract

Inference-time alignment for diffusion models aims to adapt a pre-trained diffusion model toward a target distribution without retraining the base score network, thereby preserving the generative capacity of the base model while enforcing desired properties at the inference time. A central mechanism for achieving such alignment is guidance, which modifies the sampling dynamics through an additional drift term. In this work, we introduce Doob's matching, a novel framework for guidance estimation grounded in Doob's hh-transform. Our approach formulates guidance as the gradient of logarithm of an underlying Doob's hh-function and employs gradient-penalized regression to simultaneously estimate both the hh-function and its gradient, resulting in a consistent estimator of the guidance. Theoretically, we establish non-asymptotic convergence rates for the estimated guidance. Moreover, we analyze the resulting controllable diffusion processes and prove non-asymptotic convergence guarantees for the generated distributions in the 2-Wasserstein distance.

Submission:1/10/2026
Comments:0 comments
Subjects:Data Science; Data Science
Original Source:
Was this helpful?

Discussion (0)

Please sign in to join the discussion.

No comments yet. Be the first to share your thoughts!