Flow Matching (FM) is an efficient generative modeling framework, but aligning it with human preferences remains underexplored.~Although applying Direct Preference Optimization (DPO) to diffusion models has yielded improvements, directly extending DPO-like methods to FM poses three challenges: 1) Incompatibility with ODE-based models, 2) Heavy computational cost from full model fine-tuning, and 3) Reliance on reference model quality. To address these limitations, we propose Preference Classifier for Flow Matching (PC-Flow), a novel reference-free preference alignment framework. Specifically, we reinterpret FM’s deterministic ODE as an equivalent SDE to enable DPO-style learning. Then, we introduce a lightweight classifier to model relative preferences exclusively. This approach decouples alignment from the generative model, eliminating the need for costly fine-tuning or a reference model. Theoretically, PC-Flow guarantees consistent preference-guided distribution evolution, achieves a DPO-equivalent objective without a reference model, and progressively steers generation toward preferred outputs. Experiments show that PC-Flow achieves DPO-level alignment with significantly lower training costs.

Recent advances in diffusion models focus on efficiently handling conditional generative tasks without extra training. The process involves decomposing the result into two components: 1. unconditional sample, generated in the absence of conditions; 2. condition correction, adjusting unconditional sample to include the guidance image. This adjustment is quantified by the pixel-level measure, where the latent is decoded back into a pixel image, and the forward operator translates the noisy image into the guidance domain for comparison with the guidance image. To enhance the fidelity of condition correction, we propose a learnable latent forward operator, focusing on latent-space consistency with the expectation that this latent-space consistency approximates the pixel-level fidelity measure. The encoder translates the guidance image into the latent space, and a correctional operator is proposed to rectify model mismatching in the latent guidance model. The determination of the condition term and the correction estimation is akin to solving a blind inverse problem. Our EMControl employs the Expectation-Maximization (EM) algorithm to solve the blind inverse problem during the reverse sampling process. This technique ensures that samples, once consistent with the guidance, are accurately mapped back onto the noisy data manifold, adhering to the data's inherent distribution. The EMControl has proven its effectiveness by delivering superior performance in conditional diffusion generation tasks compared to previous approaches. Moreover, its application to multiple-condition scenarios underscores its versatility and robustness across a range of generative tasks.

In training-free conditional generative tasks, diffusion models utilize differentiable loss functions to steer the generative reverse process, necessitating modifications to sampling algorithms like DDPM and DDIM. However, such adjustments likely reduce flexibility and reliability. In this paper, we propose NoiseCtrl, a sampling-algorithm-agnostic technique for controlled image generation. Essentially, diffusion models generate denoised results zt−1by adding a predicted meanµtwith random noise ϵt. NoiseCtrl specifically adjusts the random noise while leaving the underlying sampling algorithms unchanged. At each step t, NoiseCtrl converts the unconditional Gaussian noise into conditional noise $\varepsilon _t^\prime $ by substituting the isotropic Gaussian distribution with the von Mises–Fisher distribution. This substitution introduces a directional focus while preserving the randomness required for conditional image generation. Thanks to this non-intrusive design, NoiseCtrl is straightforward to integrate and has been extensively validated through experiments, demonstrating its adaptability for different diffusion algorithms and superior performance across various conditional generation tasks.

In current training-free Conditional Diffusion Models (CDM), the sampling process is steered by the gradient, which measures the discrepancy between the guidance and the condition extracted by a pre-trained condition extraction network. These methods necessitate small guidance steps, resulting in longer sampling times. To address the issue of slow sampling, we introduce AccCtr, a method that simplifies the conditional sampling algorithm by maximizing the sum of two objectives. The local maximum set of one objective is contained within the local maximum set of the other. Leveraging this relationship, we decompose the joint optimization into two parts, alternately maximizing each objective. By analyzing the steps involved in optimizing these objectives, we identify the most time-consuming steps and recommend retraining condition extraction network—a relatively simple task—to reduce its computational cost. Integrating AccCtr into current CDMs is a seamless task that does not impose a significant computational burden. Extensive testing has demonstrated that AccCtr offers superior sample quality and faster generation times.

Recent diffusion model advancements aim to handle conditional generative tasks without extra training. Existing training-free methods add a correction term at each denoising step, but they often face computational instability and lack controllability, especially with limited samples and large noise. We propose a new approach using the von Mises-Fisher (vMF) distribution to model the denoised result, turning the conditional generation task into an estimation problem for vMF parameters. We formulate the conditional diffusion model as a mean vector estimation problem for the Gaussian distribution, noting that this can be seen as an estimation problem from noisy observations. When the sampling number is small, the estimation is unstable. To address this, we optimize the mean vector of the vMF distribution by minimizing the KL divergence between the prior and posterior distributions. This approach not only addresses the computational instability but also improves the controllability and quality of the generated results. Once these parameters are determined, the denoised result can be sampled directly from the vMF distribution. Estimating the parameters requires minimal additional code and incurs negligible computational overhead while significantly improving performance. Extensive experiments across various conditional generation tasks, including depth maps, edge detection, segmentation, and style guidance, demonstrate the superiority and versatility of our method. Our approach consistently outperforms existing training-free methods and even surpasses some training-required methods in terms of visual quality and controllability.

Vision-language models (VLMs) has demonstrated impressive cross-modal alignment. However, their internal mechanisms of associating text concepts with visual patterns remain opaque. This opacity raises a critical question: What visual patterns do VLMs inherently associate with text concepts? Current methods for decoding representations of VLMs often produce suboptimal outputs, hindering to probe the clear visual patterns. To address this, we introduce Generative Semantic Probing (GSP), a novel training-free framework that synthesizes images to probe the implicit semantic preferences of VLMs. Our method generates visual patterns that maximize the similarity to the target text embeddings, through three core components: (1) Hierarchical Feature Decomposition, which decomposes the image generation across multi-scale feature levels; (2) Feature Space Constraint, which constrains the optimization within semantically meaningful feature subspace; (3) Quality Assessment Module, which ensures the generation of visually plausible outputs. Experiments validate our method's strengths in high-fidelity image generation and interpretable model analysis. Beyond text-to-image generation, style transfer and image editing applications, our framework enables unprecedented visualization of VLMs' decision boundaries. By exposing implicit preferences and systematic biases in the cross-modal association, our work provides a valuable insight for both understanding and improvement of the vision-language alignment.

Text-to-image diffusion models are typically trained on large-scale web data, often resulting in outputs that misalign with human preferences. Inspired by preference learning in large language models, we propose ABC (Alignment by Classification), a simple yet effective framework for aligning diffusion models with human preferences. In contrast to prior DPO-based methods that depend on suboptimal supervised fine-tuned (SFT) reference models, ABC assumes access to an ideal reference model perfectly aligned with human intent and reformulates alignment as a classification problem. Under this view, we recognize that preference data naturally forms a semi-supervised classification setting. To address this, we propose a data augmentation strategy that transforms preference comparisons into fully supervised training signals. We then introduce a classification-based ABC loss to guide alignment. Our alignment by classification approach could effectively steer the diffusion model toward the behavior of the ideal reference. Experiments on various diffusion models show that our ABC consistently outperforms existing baselines, offering a scalable and robust solution for preference-based text-to-image fine-tuning.

A major challenge in using diffusion models is aligning outputs with user-defined conditions. Existing conditional generation methods fall into two major categories: classifier-based guidance, which requires differentiable target models and gradient-based correction; and classifier-free guidance, which embeds conditions directly into the diffusion model but demands expensive joint training and architectural coupling. In this work, we introduce a third paradigm: DISCrete nOise (DISCO) guidance, which replaces the continuous conditional correction term with a finite codebook of discrete noise vectors sampled from a Gaussian prior. Conditional generation is reformulated as a code selection task, and we train prediction network to choose the optimal code given the intermediate diffusion state and the conditioning input. Our approach is differentiability-free, and training-efficient, avoiding the gradient computation and architectural redundancy of prior methods. Empirical results demonstrate that DISCO achieves competitive controllability while substantially reducing resource demands, positioning it as a scalable and effective alternative for conditional diffusion generation.

Anomaly detection identifies instances that deviate the distribution of the normal class. Recently, the diffusion models have shown great promise. Our research revealed that by training the diffusion model solely on normal data, it is able to transform both normal and anomalous samples into normal images. Employing this discovery, we propose ODD (One-Class Anomaly Detection via the Diffusion model), which consists of: a diffusion model to convert both normal and anomalous data into normal data, and a similarity network enhanced with outlier exposure to measure the semantic distance between the input and output of the diffusion model. If the score is low, the input is considered as an anomaly instance. The ODD is evaluated on a variety of datasets. Both qualitative and quantitative results demonstrate that our method outperforms existing state-of-the-art techniques.

Video colorization aims to add color to black-and-white films. However, propagating color information to the whole video clip accurately is a challenging task. In this paper, we propose Flow-Guided Transformer for Video Colorization (FGTVC), consisting of a Global Motion Aggregation (GMA) module, Residual modules, Flow-Guided Attention blocks (FGAB) based on encoder and decoder, to exploit the information from the neighbor patch with high similarity for each video patch colorization. Specifically, we employ Transformer to capture the long-distance dependencies between frames and learn non-local self-similarity in the frame. To overcome the shortcomings of previous optical flow-based methods, FGAB enjoys the guidance of optical flow to sample elements from spatio-temporal adjacent frames when calculating self-attention. Experiments show that the proposed FGTVC has an outstanding performance than the state-of-the-art methods. In addition, comprehensive findings demonstrate the superiority of our framework in real-world video colorization tasks.