图像生成 - arXivDaily 专题

2406.02328 2024-06-17 cs.SD eess.AS 78%

SimpleSpeech: Towards Simple and Efficient Text-to-Speech with Scalar Latent Transformer Diffusion Models

Dongchao Yang, Dingdong Wang, Haohan Guo, Xueyuan Chen, Xixin Wu, Helen Meng

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Accepted by InterSpeech 2024

2403.04430 2024-03-08 cs.LG cs.DC cs.NI 78%

On-demand Quantization for Green Federated Generative Diffusion in Mobile Edge Networks

Bingkun Lai, Jiayi He, Jiawen Kang, Gaolei Li, Minrui Xu, Tao zhang, Shengli Xie

专题命中效率与蒸馏：diffusion(title,abstract)

2401.06325 2024-01-15 stat.ML cs.LG math.OC stat.CO 78%

Faster Sampling without Isoperimetry via Diffusion-based Monte Carlo

Xunpeng Huang, Difan Zou, Hanze Dong, Yian Ma, Tong Zhang

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 54 pages

2401.04689 2024-01-10 math.ST stat.ME stat.TH 78%

Efficient estimation for ergodic diffusion processes sampled at high frequency

Michael Sørensen

专题命中效率与蒸馏：diffusion(title,abstract)

2312.07243 2023-12-13 cs.AI 78%

A Unified Sampling Framework for Solver Searching of Diffusion Probabilistic Models

Enshu Liu, Xuefei Ning, Huazhong Yang, Yu Wang

专题命中效率与蒸馏：diffusion(title,abstract)

2310.17153 2023-10-27 cs.LG stat.ME 78%

Hierarchical Semi-Implicit Variational Inference with Application to Diffusion Model Acceleration

Longlin Yu, Tianyu Xie, Yu Zhu, Tong Yang, Xiangyu Zhang, Cheng Zhang

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 25 pages, 13 figures, NeurIPS 2023

2305.20081 2023-10-27 cs.LG cs.AI 78%

Efficient Diffusion Policies for Offline Reinforcement Learning

Bingyi Kang, Xiao Ma, Chao Du, Tianyu Pang, Shuicheng Yan

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Accepted by NeurIPS 2023

2206.00646 2023-10-24 math.PR math.OC stat.ME 78%

Importance sampling for stochastic reaction-diffusion equations in the moderate deviation regime

Ioannis Gasteratos, Michael Salins, Konstantinos Spiliopoulos

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Version to appear in Stochastics and Partial Differential Equations: Analysis and Computations. 46 pages

2308.00146 2023-10-03 cs.LG cs.SI 78%

DiffusAL: Coupling Active Learning with Graph Diffusion for Label-Efficient Node Classification

Sandra Gilhuber, Julian Busch, Daniel Rotthues, Christian M. M. Frey, Thomas Seidl

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Accepted @ ECML PKDD 2023. This is the author's version of the work. The definitive Version of Record will be published in the Proceedings of ECML PKDD 2023

Journal ref ECML PKDD 2023: Machine Learning and Knowledge Discovery in Databases: Research Track pp 75-91

2308.02157 2023-08-07 cs.LG 78%

Improved Order Analysis and Design of Exponential Integrator for Diffusion Models Sampling

Qinsheng Zhang, Jiaming Song, Yongxin Chen

专题命中效率与蒸馏：diffusion(title,abstract)

2303.03509 2023-05-10 cs.AR 78%

SPARTA: Spatial Acceleration for Efficient and Scalable Horizontal Diffusion Weather Stencil Computation

Gagandeep Singh, Alireza Khodamoradi, Kristof Denolf, Jack Lo, Juan Gómez-Luna, Joseph Melber, Andra Bisca, Henk Corporaal, Onur Mutlu

专题命中效率与蒸馏：diffusion(title,abstract)

2304.04262 2023-04-11 cs.LG 78%

A Comprehensive Survey on Knowledge Distillation of Diffusion Models

Weijian Luo

专题命中效率与蒸馏：diffusion(title,abstract)

2202.12927 2023-01-26 math.AP cs.LG cs.NA math.NA math.OC 78%

A blob method for inhomogeneous diffusion with applications to multi-agent control and sampling

Katy Craig, Karthik Elamvazhuthi, Matt Haberland, Olga Turanova

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 56 pages, 8 figures

2103.15457 2022-06-24 math.ST math.PR stat.TH 78%

Estimation of ergodic square-root diffusion under high-frequency sampling

Yuzhong Cheng, Nicole Hufnagel, Hiroki Masuda

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Figure 3 is exchanged. Previously, the simulations from Figure 2 were also shown here

2009.08470 2020-12-02 astro-ph.IM astro-ph.GA 78%

AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation

Ashley Spindler, James E. Geach, Michael J. Smith

专题命中效率与蒸馏：image generation(title,abstract)

Comments Accepted in MNRAS. 23 Pages, 16 Figures. GitHub: https://github.com/AshleySpindler/AstroVaDEr-Public

2004.02313 2020-04-07 math.PR 78%

Exact simulation of diffusion first exit times: algorithm acceleration

Samuel Herrmann, Cristina Zucca

专题命中效率与蒸馏：diffusion(title,abstract)

1907.12590 2020-02-19 math.NA cs.NA 78%

Neutron transport criticality calculations using a parallel monolithic multilevel Schwarz preconditioner together with a nonlinear diffusion acceleration method

Fande Kong

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Changed title to better represent the content

1902.08731 2019-02-26 physics.comp-ph 78%

Nonlinear Diffusion Acceleration of the Least-Squares Transport Equation in Geometries with Voids

Hans Hammer, Jim Morel, Yaqi Wang

专题命中效率与蒸馏：diffusion(title,abstract)

1807.09637 2019-02-01 eess.SP 78%

Efficient sampling and robust 3D diffusion magnetic resonance imaging signal reconstruction

Alice P. Bates, Zubair Khalid, Jason D. McEwen, Rodney A. Kennedy, Alessandro Daducci, Erick J. Canales-Rodríguez

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 19 pages, with 6 pages supplementary material attached at the end, submitted to PLOS ONE

1704.01055 2018-06-26 math.ST stat.TH 78%

One-step Local M-estimator for Integrated Jump-Diffusion Models

Yuping Song, Hanchao Wang

专题命中效率与蒸馏：diffusion(title,abstract)

Comments There are some lackness for simulation study and empirical analysis, some error correction to be done for the detailed proof

1806.03874 2018-06-12 physics.med-ph 78%

Integrated and efficient diffusion-relaxometry using ZEBRA

Jana Hutter, Paddy J Slator, Daan C Christiaens, Rui Pedro AG Teixeira, Thomas Roberts, Laurence Jackson, Anthony N Price, Shaihan Malik, Joseph V Hajnal

专题命中效率与蒸馏：diffusion(title,abstract)

1706.06682 2017-09-26 physics.med-ph cs.CG math.OC 78%

Single- and Multiple-Shell Uniform Sampling Schemes for Diffusion MRI Using Spherical Codes

Jian Cheng, Dinggang Shen, Pew-Thian Yap, Peter J. Basser

专题命中效率与蒸馏：diffusion(title,abstract)

Comments Accepted by IEEE transactions on Medical Imaging. Codes have been released in dmritool https://diffusionmritool.github.io/tutorial_qspacesampling.html

1702.07056 2017-02-24 cs.DM 78%

Multi-shell Sampling Scheme with Accurate and Efficient Transforms for Diffusion MRI

Alice P. Bates, Zubair Khalid, Rodney A. Kennedy, Jason D. McEwen

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 1 page, 1 figure, presented as a poster at the 2017 Biomedical and Astronomical Image Processing (BASP) Workshop

1407.2261 2015-06-22 astro-ph.HE 78%

Simulations of Ion Acceleration at Non-relativistic Shocks. III. Particle Diffusion

Damiano Caprioli, Anatoly Spitkovsky

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 8 pages, 7 figures, Minor changes reflecting the version accepted to ApJ

1304.2531 2015-05-25 math.PR 78%

Recursive marginal quantization of the Euler scheme of a diffusion process

Gilles Pagès, Abass Sagna

专题命中效率与蒸馏：diffusion(title,abstract)

Comments 29 pages

2607.24157 2026-07-28 cs.CV 新提交 77%

UniGen-AR: Unifying Visual Generation with Auto-Regressive Modeling

UniGen-AR：通过自回归建模统一视觉生成

Zhipeng Bao, Zhen Zhu, Nupur Kumari, Anurag Bagchi, Yu-Xiong Wang, Pavel Tokmakov, Martial Hebert

机构 * Carnegie Mellon University（卡内基梅隆大学）； University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳 - 香槟分校）； Toyota Research Institute（丰田研究院）

专题命中效率与蒸馏：image generation(abstract)；text-to-image(abstract)；diffusion(abstract)；分类 cs.CV

AI总结研究统一视觉生成问题，提出UniGen-AR框架，将通用多模态语言模型与视觉自回归解码器配对，能为多任务生成图像值输出，相比基于扩散的基线，推理延迟低达19倍，确立视觉自回归建模为统一视觉生成的高效主干。

2605.26108 2026-06-09 cs.CV 版本更新 77%

Reinforcing Few-step Generators via Reward-Tilted Distribution Matching

通过奖励倾斜分布匹配增强少步生成器

Yushi Huang, Xiangxin Zhou, Ruoyu Wang, Chi Zhang, Jun Zhang, Tianyu Pang

机构 * Tencent Hunyuan（腾讯文英）； Hong Kong University of Science and Technology（香港科技大学）； Westlake University（西湖大学）

专题命中效率与蒸馏：image generation(abstract)；text-to-image(abstract)；diffusion(abstract)；分类 cs.CV

AI总结提出奖励倾斜分布匹配蒸馏（RTDMD）两阶段框架，结合分布匹配蒸馏与奖励引导强化学习，在仅4步推理下实现文本到图像生成的最新性能。

Comments Code and models are available at https://github.com/Harahan/RTDMD

2507.09574 2026-05-29 cs.CV cs.AI cs.CL 77%

MENTOR: Efficient Multimodal-Conditioned Tuning for Autoregressive Vision Generation Models

MENTOR: 面向自回归视觉生成模型的高效多模态条件微调

Haozhe Zhao, Zefan Cai, Shuzheng Si, Liang Chen, Jiuxiang Gu, Wen Xiao, Minjia Zhang, Junjie Hu

机构 * University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）； University of Wisconsin-Madison（威斯康星大学麦迪逊分校）； Tsinghua University（清华大学）； Peking University（北京大学）； Microsoft（微软公司）

专题命中效率与蒸馏：image generation(abstract)；text-to-image(abstract)；diffusion(abstract)；分类 cs.CV

AI总结提出MENTOR框架，通过两阶段训练范式实现自回归图像生成器与多模态输入的细粒度token级对齐，无需辅助适配器或交叉注意力模块，在DreamBench++上取得优异性能。

Comments Findings of ACL 2026

2603.06165 2026-04-28 cs.CV cs.AI 77%

Reflective Flow Sampling Enhancement

反射流采样增强

Zikai Zhou, Muyao Wang, Shitong Shao, Lichen Bai, Haoyi Xiong, Bo Han, Zeke Xie

机构 * The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））； The University of Tokyo（东京大学）； Microsoft（微软）； Hong Kong Baptist University（香港 Baptist大学）

专题命中效率与蒸馏：image generation(abstract)；text-to-image(abstract)；diffusion(abstract)；分类 cs.CV

AI总结本文提出Reflective Flow Sampling，一种针对流模型的无训练增强框架，提升文本图像生成质量与提示对齐。

2511.22177 2026-04-28 cs.LG cs.CV 77%

Designing Instance-Level Sampling Schedules via REINFORCE with James-Stein Shrinkage

通过REINFORCE与James-Stein收缩设计实例级采样计划

Peiyu Yu, Suraj Kothawade, Sirui Xie, Ying Nian Wu, Hongliang Fei

机构 * Google（谷歌）； Google DeepMind（谷歌DeepMind）； UCLA（加州大学洛杉矶分校）

专题命中效率与蒸馏：diffusion(abstract,abstract_cn)；text-to-image(abstract)；分类 cs.CV

AI总结本文提出通过REINFORCE学习实例级采样计划，结合James-Stein收缩估计器提升高维策略学习的梯度估计精度，实现文本图像对齐和生成质量提升。

Comments CVPR 2026; 23 pages

1. 效率与蒸馏 2537 篇

SimpleSpeech: Towards Simple and Efficient Text-to-Speech with Scalar Latent Transformer Diffusion Models

On-demand Quantization for Green Federated Generative Diffusion in Mobile Edge Networks

Faster Sampling without Isoperimetry via Diffusion-based Monte Carlo

Efficient estimation for ergodic diffusion processes sampled at high frequency

A Unified Sampling Framework for Solver Searching of Diffusion Probabilistic Models

Hierarchical Semi-Implicit Variational Inference with Application to Diffusion Model Acceleration

Efficient Diffusion Policies for Offline Reinforcement Learning

Importance sampling for stochastic reaction-diffusion equations in the moderate deviation regime

DiffusAL: Coupling Active Learning with Graph Diffusion for Label-Efficient Node Classification

Improved Order Analysis and Design of Exponential Integrator for Diffusion Models Sampling

SPARTA: Spatial Acceleration for Efficient and Scalable Horizontal Diffusion Weather Stencil Computation

A Comprehensive Survey on Knowledge Distillation of Diffusion Models

A blob method for inhomogeneous diffusion with applications to multi-agent control and sampling

Estimation of ergodic square-root diffusion under high-frequency sampling

AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation

Exact simulation of diffusion first exit times: algorithm acceleration

Neutron transport criticality calculations using a parallel monolithic multilevel Schwarz preconditioner together with a nonlinear diffusion acceleration method

Nonlinear Diffusion Acceleration of the Least-Squares Transport Equation in Geometries with Voids

Efficient sampling and robust 3D diffusion magnetic resonance imaging signal reconstruction

One-step Local M-estimator for Integrated Jump-Diffusion Models

Integrated and efficient diffusion-relaxometry using ZEBRA

Single- and Multiple-Shell Uniform Sampling Schemes for Diffusion MRI Using Spherical Codes

Multi-shell Sampling Scheme with Accurate and Efficient Transforms for Diffusion MRI

Simulations of Ion Acceleration at Non-relativistic Shocks. III. Particle Diffusion

Recursive marginal quantization of the Euler scheme of a diffusion process

UniGen-AR: Unifying Visual Generation with Auto-Regressive Modeling

Reinforcing Few-step Generators via Reward-Tilted Distribution Matching

MENTOR: Efficient Multimodal-Conditioned Tuning for Autoregressive Vision Generation Models

Reflective Flow Sampling Enhancement

Designing Instance-Level Sampling Schedules via REINFORCE with James-Stein Shrinkage