In this paper, we introduce the Diff-Instruct* (DI*), an image data-free approach for building one-step text-to-image generative models that align with human preference while maintaining the ability to generate highly realistic images. We frame human preference alignment as online reinforcement learning using human feedback (RLHF), where the goal is to maximize the reward function while regularizing the generator distribution to remain close to a reference diffusion process. Unlike traditional RLHF approaches, which rely on the KL divergence for regularization, we introduce a novel score-based divergence regularization, which leads to significantly better performances. Although the direct calculation of this preference alignment objective remains intractable, we demonstrate that we can efficiently compute its gradient by deriving an equivalent yet tractable loss function. Remarkably, we used Diff-Instruct* to train a Stable Diffusion-XL-based 1-step model, the 2.6B DI*-SDXL-1step text-to-image model, which can generate images of a resolution of 1024x1024 with only 1 generation step. DI*-SDXL-1step model uses only 1.88% inference time and 29.30% GPU memory cost to outperform 12B FLUX-dev-50step significantly in PickScore, ImageReward, and CLIPScore on Parti prompt benchmark and HPSv2.1 on Human Preference Score benchmark, establishing a new state-of-the-art benchmark of human-preferred 1-step text-to-image generative models. Besides the strong quantitative performances, extensive qualitative comparisons also confirm the advantages of DI* in terms of maintaining diversity, improving image layouts, and enhancing aesthetic colors. We have released our industry-ready model on the homepage: \url{https://github.com/pkulwj1994/diff_instruct_star}.

Given an intractable distribution $p$, the problem of variational inference (VI) is to compute the best approximation $q$ from some more tractable family $\mathcal{Q}$. Most commonly the approximation is found by minimizing a Kullback-Leibler (KL) divergence. However, there exist other valid choices of divergences, and when $\mathcal{Q}$ does not contain~$p$, each divergence champions a different solution. We analyze how the choice of divergence affects the outcome of VI when a Gaussian with a dense covariance matrix is approximated by a Gaussian with a diagonal covariance matrix. In this setting we show that different divergences can be \textit{ordered} by the amount that their variational approximations misestimate various measures of uncertainty, such as the variance, precision, and entropy. We also derive an impossibility theorem showing that no two of these measures can be simultaneously matched by a factorized approximation; hence, the choice of divergence informs which measure, if any, is correctly estimated. Our analysis covers the KL divergence, the R\'enyi divergences, and a score-based divergence that compares $\nabla\log p$ and $\nabla\log q$. We empirically evaluate whether these orderings hold when VI is used to approximate non-Gaussian distributions.

Most leading implementations of black-box variational inference (BBVI) are based on optimizing a stochastic evidence lower bound (ELBO). But such approaches to BBVI often converge slowly due to the high variance of their gradient estimates. In this work, we propose batch and match (BaM), an alternative approach to BBVI based on a score-based divergence. Notably, this score-based divergence can be optimized by a closed-form proximal update for Gaussian variational families with full covariance matrices. We analyze the convergence of BaM when the target distribution is Gaussian, and we prove that in the limit of infinite batch size the variational parameter updates converge exponentially quickly to the target mean and covariance. We also evaluate the performance of BaM on Gaussian and non-Gaussian target distributions that arise from posterior inference in hierarchical and deep generative models. In these experiments, we find that BaM typically converges in fewer (and sometimes significantly fewer) gradient evaluations than leading implementations of BBVI based on ELBO maximization.