图像生成 - arXivDaily 专题

2306.02708 2025-10-10 q-fin.MF math.PR 50%

Efficient simulation of a new class of Volterra-type SDEs

Ofelia Bonesini, Giorgia Callegaro, Martino Grasselli, Gilles Pagès

专题命中效率与蒸馏：diffusion(abstract)

Comments Previously appearing as "From elephant to goldfish (and back): memory in stochastic Volterra processes"

2510.05984 2025-10-08 cs.SD cs.AI eess.AS 50%

ECTSpeech: Enhancing Efficient Speech Synthesis via Easy Consistency Tuning

Tao Zhu, Yinfeng Yu, Liejun Wang, Fuchun Sun, Wendong Zheng

机构 * School of Computer Science and Technology, Xinjiang University（新疆大学计算机科学与技术学院）； Joint International Research Laboratory of Silk Road Multilingual Cognitive Computing（丝绸之路多语种认知计算联合国际实验室）； Tsinghua University（清华大学）； Tianjin University of Technology（天津工业大学）

专题命中效率与蒸馏：diffusion(abstract)

Comments Accepted for publication by Proceedings of the 2025 ACM Multimedia Asia Conference(MMAsia '25)

2509.03091 2025-10-08 astro-ph.HE 50%

Diffusive shock acceleration: non-classical model of cosmic ray transport

A. A. Lagutin

专题命中效率与蒸馏：diffusion(abstract)

Comments 17 pages, the 5th International Symposium on Cosmic Rays and Astrophysics (ISCRA-2025)

2510.04635 2025-10-07 astro-ph.HE hep-ph 50%

The Role of Acoustic Instability in Cosmic-Ray Self-Confinement

Antonio Capanema, Pasquale Blasi, Emanuele Sobacchi

专题命中效率与蒸馏：diffusion(abstract)

Comments Presented at the 39th International Cosmic Ray Conference (ICRC 2025). 8 pages, 5 figures

2506.11893 2025-10-07 cs.LG 50%

Measurement-Aligned Sampling for Inverse Problem

Shaorong Zhang, Rob Brekelmans, Yunshu Wu, Greg Ver Steeg

机构 * University of California Riverside（加州大学河滨分校）； Vector Institute（向量研究所）

专题命中效率与蒸馏：diffusion(abstract)

2510.01790 2025-10-03 math.NA cs.NA math.AP math.DG 50%

Efficient manifold evolution algorithm using adaptive B-Spline interpolation

Muhammad Ammad, Leevan Ling

专题命中效率与蒸馏：diffusion(abstract)

2508.01622 2025-10-03 cs.RO cs.LG 50%

VFP: Variational Flow-Matching Policy for Multi-Modal Robot Manipulation

Xuanran Zhai, Qianyou Zhao, Qiaojun Yu, Ce Hao

机构 * National University of Singapore（新加坡国立大学）； Shanghai Jiao Tong University（上海交通大学）； Shanghai AI Lab（上海人工智能实验室）

专题命中效率与蒸馏：diffusion(abstract)

2410.02367 2025-10-02 cs.LG 50%

SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration

Jintao Zhang, Jia Wei, Haofeng Huang, Pengle Zhang, Jun Zhu, Jianfei Chen

机构 * Dept. of Comp. Sci. & Tech.（计算机科学与技术系）； Institute for AI（人工智能研究院）； BNRist Center（BNRist中心）； Tsinghua-Bosch Joint ML Center（清华大学-博世联合机器学习中心）； THBI Lab（THBI实验室）； Tsinghua University（清华大学）

专题命中效率与蒸馏：image generation(abstract)

Comments @inproceedings{zhang2025sageattention, title={SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration}, author={Zhang, Jintao and Wei, Jia and Zhang, Pengle and Zhu, Jun and Chen, Jianfei}, booktitle={International Conference on Learning Representations (ICLR)}, year={2025} }

Journal ref The Thirteenth International Conference on Learning Representations (ICLR 2025)

2510.00722 2025-10-02 math.NA cs.NA 50%

Carleman Linearization of Parabolic PDEs: Well-posedness, convergence, and efficient numerical methods

Bernhard Heinzelreiter, John W. Pearson

专题命中效率与蒸馏：diffusion(abstract)

Comments 32 pages

2506.05526 2025-10-01 cs.LG stat.ML 50%

On Fitting Flow Models with Large Sinkhorn Couplings

Stephen Zhang, Alireza Mousavi-Hosseini, Michal Klein, Marco Cuturi

机构 * Apple（苹果公司）； University of Toronto（多伦多大学）； University of Melbourne（墨尔本大学）

专题命中效率与蒸馏：image generation(abstract)

Comments 23 pages, 14 figures

2509.26488 2025-10-01 cs.CL 50%

dParallel: Learnable Parallel Decoding for dLLMs

Zigeng Chen, Gongfan Fang, Xinyin Ma, Ruonan Yu, Xinchao Wang

机构 * National University of Singapore（新加坡国立大学）

专题命中效率与蒸馏：diffusion(abstract)

Comments Working in progress, code base: https://github.com/czg1225/dParallel

2509.23299 2025-10-01 cs.SD eess.AS 50%

MeanFlowSE: One-Step Generative Speech Enhancement via MeanFlow

Yike Zhu, Boyi Kang, Ziqian Wang, Xingchen Li, Zihan Zhang, Wenjie Li, Longshuai Xiao, Wei Xue, Lei Xie

机构 * The Hong Kong University of Science and Technology, Hong Kong, China（香港科学与技术大学，香港，中国）； Huawei Technologies, China（华为技术有限公司，中国）

专题命中效率与蒸馏：diffusion(abstract)

Comments Submitted to ICASSP 2026

2504.10188 2025-09-30 cs.LG cs.AI 50%

Efficient Generative Model Training via Embedded Representation Warmup

Deyuan Liu, Peng Sun, Xufeng Li, Tao Lin

机构 * Westlake University（西湖大学）； Zhejiang University（浙江大学）； Nanjing University（南京大学）

专题命中效率与蒸馏：diffusion(abstract)

2407.00611 2025-09-30 cs.DC 50%

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

Ziming Liu, Shaoyu Wang, Shenggan Cheng, Zhongkai Zhao, Kai Wang, Xuanlei Zhao, James Demmel, Yang You

专题命中效率与蒸馏：diffusion(abstract)

2507.18578 2025-09-29 cs.CL 50%

Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs

Feng Hong, Geng Yu, Yushi Ye, Haicheng Huang, Huangjie Zheng, Ya Zhang, Yanfeng Wang, Jiangchao Yao

机构 * Cooperative Medianet Innovation Center, Shanghai Jiao Tong University（上海交通大学合作中位网创新中心）； Apple（苹果公司）； School of Artificial Intelligence, Shanghai Jiao Tong University（上海交通大学人工智能学院）

专题命中效率与蒸馏：diffusion(abstract)

2509.20744 2025-09-26 cs.AI 50%

Parallel Thinking, Sequential Answering: Bridging NAR and AR for Efficient Reasoning

Qihang Ai, Haiyun Jiang

机构 * Nanyang Technological University（南洋理工大学）； Shanghai Jiao Tong University（上海交通大学）

专题命中效率与蒸馏：diffusion(abstract)

Comments 4 pages

2509.19816 2025-09-25 cs.LG 50%

An Efficient Conditional Score-based Filter for High Dimensional Nonlinear Filtering Problems

Zhijun Zeng, Weiye Gan, Junqing Chen, Zuoqiang Shi

机构 * Department of Mathematical Sciences, Tsinghua University（清华大学数学科学系）； Yau Mathematical Sciences Center, Tsinghua University（清华大学叶中伟数学科学中心）； Yanqi Lake Beijing Institute of Mathematical Sciences and Applications（北京燕琦湖数学科学与应用研究所）

专题命中效率与蒸馏：diffusion(abstract)

2509.17894 2025-09-24 cs.LG 50%

Optimizing Inference in Transformer-Based Models: A Multi-Method Benchmark

Siu Hang Ho, Prasad Ganesan, Nguyen Duong, Daniel Schlabig

专题命中效率与蒸馏：diffusion(abstract)

Comments 6 pages, 4 figures. Technical report

2509.14847 2025-09-23 math.NA cs.NA 50%

A class of flexible and efficient partitioned Runge-Kutta-Chebyshev methods for some time-dependent partial differential equations

Xiao Tang, Junwei Huang

专题命中效率与蒸馏：diffusion(abstract)

2509.11284 2025-09-16 cs.LG physics.comp-ph 50%

PINGS: Physics-Informed Neural Network for Fast Generative Sampling

Achmad Ardani Prasha, Clavino Ourizqi Rachmadi, Muhamad Fauzan Ibnu Syahlan, Naufal Rahfi Anugerah, Nanda Garin Raditya, Putri Amelia, Sabrina Laila Mutiara, Hilman Syachr Ramadhan

机构 * Faculty of Computer Science, Universitas Mercu Buana（计算机科学学院，默克鲁巴纳大学）； Alumni, Universitas Mercu Buana（默克鲁巴纳大学校友）

专题命中效率与蒸馏：diffusion(abstract)

Comments 19 pages, 4 figures

2505.20514 2025-09-16 hep-th math-ph math.MP 50%

Spontaneous quantization of the Yang--Mills gradient flow

Alexander Migdal

专题命中效率与蒸馏：diffusion(abstract)

Comments 26 pages, 5 figures, This is the final revision submitted to NPB, with added discussion of previous analysis of minimal surfaces as solution of the loop equation

2509.09748 2025-09-15 cs.SD eess.AS 50%

DiTReducio: A Training-Free Acceleration for DiT-Based TTS via Progressive Calibration

Yanru Huo, Ziyue Jiang, Zuoli Tang, Qingyang Hong, Zhou Zhao

机构 * Zhejiang University（浙江大学）； Xiamen University（厦门大学）； Wuhan University（武汉大学）

专题命中效率与蒸馏：diffusion(abstract)

2509.05001 2025-09-08 math.NA cs.NA 50%

Synthetic Acceleration Preconditioners for Parametric Radiative Transfer Equations based on Trajectory-Aware Reduced Order Models

Ning Tang, Zhichao Peng

专题命中效率与蒸馏：diffusion(abstract)

Comments 32 pages, 9 figures

2509.04996 2025-09-08 cs.RO 50%

FLOWER: Democratizing Generalist Robot Policies with Efficient Vision-Language-Action Flow Policies

Moritz Reuss, Hongyi Zhou, Marcel Rühle, Ömer Erdinç Yağmurlu, Fabian Otto, Rudolf Lioutikov

机构 * Intuitive Robots Lab（直观机器人实验室）； Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）； Microsoft Research（微软研究院）

专题命中效率与蒸馏：diffusion(abstract)

Comments Published at CoRL 2025

2509.04920 2025-09-08 math.PR 50%

Efficient estimation of jump parameters for stochastic differential equations driven by L{é}vy processes

Elise Bayraktar, Emmanuelle Clément

专题命中效率与蒸馏：diffusion(abstract)

2407.15161 2025-09-05 cs.RO cs.AI cs.LG 50%

FFHFlow: Diverse and Uncertainty-Aware Dexterous Grasp Generation via Flow Variational Inference

Qian Feng, Jianxiang Feng, Zhaopeng Chen, Rudolph Triebel, Alois Knoll

机构 * School of Information Computation and Technology, Technical University of Munich (TUM)（技术大学慕尼黑信息计算技术学院）； Institute of Robotics and Mechatronics, German Aerospace Center (DLR)（德国航空航天中心机器人与机电研究所）； Department of Informatics, Karlsruhe Institute of Technology (KIT)（卡尔斯鲁厄理工学院信息学院）

专题命中效率与蒸馏：diffusion(abstract)

Comments First two authors contributed equally, whose ordering decided via coin-tossing. Accepted for CoRL 2025

2509.02607 2025-09-04 eess.IV cs.AI 50%

Towards Digital Twins for Optimal Radioembolization

Nisanth Kumar Panneerselvam, Guneet Mummaneni, Emilie Roncali

专题命中效率与蒸馏：diffusion(abstract)

2509.02069 2025-09-03 cs.LG q-bio.QM 50%

Data-Dependent Smoothing for Protein Discovery with Walk-Jump Sampling

Srinivas Anumasa, Barath Chandran. C, Tingting Chen, Dianbo Liu

机构 * National University of Singapore（新加坡国立大学）； Indian Institute of Technology, Roorkee（印度理工学院罗尔克学院）

专题命中效率与蒸馏：diffusion(abstract)

2509.01942 2025-09-03 stat.AP gr-qc 50%

Efficient Bayesian Sampling with Langevin Birth-Death Dynamics

Alex Leviyev, Francesco Iacovelli, Aaron Zimmerman

专题命中效率与蒸馏：diffusion(abstract)

2509.01679 2025-09-03 cs.LG cs.NA math.NA stat.ML 50%

Efficient Transformer-Inspired Variants of Physics-Informed Deep Operator Networks

Zhi-Feng Wei, Wenqian Chen, Panos Stinis

机构 * Pacific Northwest National Laboratory（太平洋西北国家实验室）； University of Washington（华盛顿大学）； Brown University（布朗大学）

专题命中效率与蒸馏：diffusion(abstract)

Comments Code will be released upon acceptance

1. 效率与蒸馏 2557 篇

Efficient simulation of a new class of Volterra-type SDEs

ECTSpeech: Enhancing Efficient Speech Synthesis via Easy Consistency Tuning

Diffusive shock acceleration: non-classical model of cosmic ray transport

The Role of Acoustic Instability in Cosmic-Ray Self-Confinement

Measurement-Aligned Sampling for Inverse Problem

Efficient manifold evolution algorithm using adaptive B-Spline interpolation

VFP: Variational Flow-Matching Policy for Multi-Modal Robot Manipulation

SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration

Carleman Linearization of Parabolic PDEs: Well-posedness, convergence, and efficient numerical methods

On Fitting Flow Models with Large Sinkhorn Couplings

dParallel: Learnable Parallel Decoding for dLLMs

MeanFlowSE: One-Step Generative Speech Enhancement via MeanFlow

Efficient Generative Model Training via Embedded Representation Warmup

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs

Parallel Thinking, Sequential Answering: Bridging NAR and AR for Efficient Reasoning

An Efficient Conditional Score-based Filter for High Dimensional Nonlinear Filtering Problems

Optimizing Inference in Transformer-Based Models: A Multi-Method Benchmark

A class of flexible and efficient partitioned Runge-Kutta-Chebyshev methods for some time-dependent partial differential equations

PINGS: Physics-Informed Neural Network for Fast Generative Sampling

Spontaneous quantization of the Yang--Mills gradient flow

DiTReducio: A Training-Free Acceleration for DiT-Based TTS via Progressive Calibration

Synthetic Acceleration Preconditioners for Parametric Radiative Transfer Equations based on Trajectory-Aware Reduced Order Models

FLOWER: Democratizing Generalist Robot Policies with Efficient Vision-Language-Action Flow Policies

Efficient estimation of jump parameters for stochastic differential equations driven by L{é}vy processes

FFHFlow: Diverse and Uncertainty-Aware Dexterous Grasp Generation via Flow Variational Inference

Towards Digital Twins for Optimal Radioembolization

Data-Dependent Smoothing for Protein Discovery with Walk-Jump Sampling

Efficient Bayesian Sampling with Langevin Birth-Death Dynamics

Efficient Transformer-Inspired Variants of Physics-Informed Deep Operator Networks