The paper introduces a theoretical framework to study how perturbations introduced during diffusion model sampling affect the generated output. It defines a path cost based on the Kullback–Leibler divergence between perturbed and reference trajectories, showing that this cost bounds but does not fully determine output changes. Using a response identity, the authors link local perturbation propagation to changes in feature means, revealing that first‑order responses can miss significant output variations. Experiments on pretrained diffusion models demonstrate varying output sensitivity across sampling stages and spatial frequencies, and the framework successfully identifies problematic sampling intervals in cache‑based acceleration.
By Wei Guo, Yaowen Zhang, Xingtong Ge, Jun Zhang
arXiv:2608. 13043v1 Announce Type: new Abstract: Diffusion models have achieved dominant performance in visual generation but suffer from substantial inference overhead.
By Xichen Ye, Yifan Wu, Zhikang Xie, Xiangyu Yue, Cheng Jin, Weizhong Zhang
arXiv:2506. 13058v2 Announce Type: replace-cross Abstract: Diffusion probabilistic models (DPMs) have demonstrated remarkable success in visual generation.
By Hu Yu, Hao Luo, Xueyang Fu, Jie Huang, Fan Wang, Feng Zhao
arXiv:2602. 02908v2 Announce Type: replace-cross Abstract: Diffusion models trained on different, non-overlapping subsets of a dataset often produce strikingly similar outputs when given the same noise seed.
By Binxu Wang, Jacob Zavatone-Veth, Cengiz Pehlevan
The paper studies how to allocate a fixed computational budget across the denoising steps of diffusion models to improve sample quality at deployment. It shows that the expected benefit of evaluating multiple candidates at a step can be decomposed into a step‑specific sensitivity and a universal sample‑size factor, and that the optimal allocation follows a water‑filling structure. Experiments demonstrate that this allocation achieves the same quality as a uniform strategy while reducing function evaluations by 20–50%.
By Yuan Cao, Yifu Tang, Hangqi Li, Zeyu Zheng
arXiv:2609.39648v1 Announce Type: cross
Abstract: Diffusion models are typically viewed as stochastic processes that transform noise into data. We take a complementary perspective: a diffusion model...
By Cristina L\'opez Amado, Marco Fumero, Francesco Locatello
Understanding the mechanism behind why diffusion models memorize their training data remains an open question in generative modeling. In this paper, we examine this question from two perspectives: (a)...
The paper investigates how repeatedly training score-based diffusion models on synthetic data can cause their outputs to drift from the target distribution. It provides theoretical upper and lower bounds on the divergence accumulated over successive training rounds, distinguishing regimes based on score estimation error and the fraction of fresh data used. Empirical experiments on synthetic data and images support the theoretical findings.
By Nail B. Khelifa, Richard E. Turner, Ramji Venkataramanan
Accelerating Video Diffusion via Training-Free Trajectory Routing (TRACK) introduces a heterogeneous denoising strategy that switches between large and small diffusion models at selected steps, determined by a calibration process that measures disagreement between model predictions. By routing quality-sensitive steps to the large model and low-disagreement steps to the small model, TRACK achieves significant speedups—up to 2.73×—across several video diffusion benchmarks while maintaining comparable quality and diversity. The method requires no retraining, architectural changes, or online dual-model evaluation, making it a practical acceleration paradigm for video diffusion.
arXiv:2602.24208v2 Announce Type: replace-cross
Abstract: Diffusion models achieve state-of-the-art video generation quality, but their inference remains expensive due to the large number of sequenti...
By Yasaman Haghighi, Alexandre Alahi
The paper introduces a data‑free learning objective called relative trajectory balance for training diffusion models to sample from a posterior defined by a diffusion prior and an arbitrary black‑box constraint or likelihood. It proves asymptotic correctness of this objective and demonstrates its use across vision, language, and multimodal tasks, including classifier guidance, language infilling, and text‑to‑image generation. Additionally, the method is applied to continuous control with a score‑based behavior prior, achieving state‑of‑the‑art results in offline reinforcement learning.
By Siddarth Venkatraman, Moksh Jain, Luca Scimeca, Minsu Kim, Marcin Sendera, Mohsin Hasan, Luke Rowe, Sarthak Mittal, Pablo Lemos, Emmanuel Bengio, Alexandre Adam, Jarrid Rector-Brooks, Yoshua Bengio, Glen Berseth, Esmeralda S. Whitammer
arXiv:2506. 11378v3 Announce Type: replace Abstract: Sampling in score-based diffusion models can be performed by solving either a reverse-time stochastic differential equation (SDE) parameterized by an arbitrary stochasticity function or a probability flow ODE, corresponding to setting this stochasticity function to zero.
By Bernardo P. Schaeffer, Ricardo M. S. Rosa, Glauco Valle