홈 › Papers

Improved Distributional Diffusion Models

2026-09-29 · Tommaso Martorella, Alexandre Galashov, Felix Krause, Stefan Andreas Baumann, Valentin De Bortoli, Arthur Gretton, Björn Ommer hf

Distributional Diffusion Models (DDMs) replace the standard mean-prediction denoiser with a distributional denoiser trained via a scoring rule objective, learning a stochastic approximation to p(x_1 mid x_t) rather than its conditional mean. However, scaling DDMs to modern image-generation settings faces two obstacles: (i) multi-particle training incurs overhead that scales with the number of particles, (ii) DDMs use globally fixed scoring rule hyperparameters, forcing a single trade-off across sampling budgets. We mitigate these limitations by deferring particle expansion to late transformer layers, and the hyperparameter trade-off by introducing time-dependent scoring rule schedules informed by the dynamical regimes of~Biroli2024. Combined with a DiT-based latent setup, these changes make DDM training practical on class-conditional ImageNet-256^2, achieving 4.48 FID at 4 steps and 2.38 at 50 steps with DiT-XL/2, from a single model trained from scratch in one stage, without a teacher, self-distillation or JVPs. The result is a stochastic few-step generator whose FID does not degrade as the sampling budget grows from 4 to 50 NFE, and the same recipe transfers to text-to-image generation. Code and pre-trained models available at https://github.com/CompVis/iDDM.

📄 PDF Abstract BibTeX arXiv:2609.37147

Code (0)

등록된 구현이 없습니다.

Tasks

Text-to-Image Generation

Regularized Personalization of Text-to-Image Diffusion Models without Distributional Drift

2025-05-26 · Gihoon Kim, Hyungjin Park, Taesup Kim

Personalization using text-to-image diffusion models involves adapting a pretrained model to novel subjects with only a few image examples. This task presents a fundamental challenge, as the model must not only learn the…

Distributional Reinforcement Learning with Diffusion Bridge Critics

2026-02-05 · Shutong Ding, Yimiao Zhou, Ke Hu, Mokai Pan 외 arxiv

Recent advances in diffusion-based reinforcement learning (RL) methods have demonstrated promising results in a wide range of continuous control tasks. However, existing works in this field focus on the application of di…

Reinforcement Learning Continuous Control

Diffusion Reward Models

2026-09-27 · Xiangyang Wang, Bingxiang He, Zeyuan Liu, Jiaze WangZiqing Qiao 외 hf

Reward models underpin the alignment of large language models, yet the dominant designs reduce each prompt--response pair to a point estimate or to a distribution from a fixed parametric family. This is at odds with huma…

Density Estimation

Distributional Soft Actor-Critic with Diffusion Policy

2025-07-02 · Tong Liu, Yinuo Wang, Xujie Song, Wenjun Zou 외 arxiv

Reinforcement learning has been proven to be highly effective in handling complex control tasks. Traditional methods typically use unimodal distributions, such as Gaussian distributions, to model the output of value dist…

Reinforcement Learning

DS-Diffusion: Data Style-Guided Diffusion Model for Time-Series Generation

2025-09-23 · Mingchun Sun, Rongqiang Zhao, Hengrui Hu, Songyu Ding 외 arxiv

Diffusion models are the mainstream approach for time series generation tasks. However, existing diffusion models for time series generation require retraining the entire framework to introduce specific conditional guida…