语言大模型 / LLM - arXivDaily 专题

2510.00373 2025-10-02 cs.LG cs.AI cs.NE cs.SY eess.SY 90%

Combining Large Language Models and Gradient-Free Optimization for Automatic Control Policy Synthesis

Carlo Bosio, Matteo Guarrera, Alberto Sangiovanni-Vincentelli, Mark W. Mueller

机构 * UC Berkeley, Department of Mechanical Engineering（伯克利大学机械工程系）

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)；分类 cs.AI、cs.LG

Comments 8 pages, 7 figures

2510.00490 2025-10-02 cs.CR 89%

Has the Two-Decade-Old Prophecy Come True? Artificial Bad Intelligence Triggered by Merely a Single-Bit Flip in Large Language Models

Yu Yan, Siqi Lu, Yang Gao, Zhaoxuan Li, Ziming Zhao, Qingjun Yuan, Yongjuan Wang

专题命中效率与部署：large language model(title,abstract)；language model(title,abstract)；LLM(abstract)

Comments 19 pages

2510.00028 2025-10-02 cs.LG cs.AI 88%

Rethinking RoPE Scaling in Quantized LLM: Theory, Outlier, and Channel-Band Analysis with Weight Rescaling

Ye Qiao, Haocheng Xu, Xiaofan Zhang, Sitao Huang

机构 * Department of EECS, University of California, Irvine, USA（加州大学尔湾分校电子工程与计算机科学系）； Google（谷歌）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；post-training(abstract)

2510.00555 2025-10-02 cs.HC cs.AI 87%

PromptPilot: Improving Human-AI Collaboration Through LLM-Enhanced Prompt Engineering

Niklas Gutheil, Valentin Mayer, Leopold Müller, Jörg Rommelt, Niklas Kühl

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；prompting(abstract)

Comments Preprint version. Accepted for presentation at the International Conference on Information Systems (ICIS 2025). Please cite the published version when available

2510.00515 2025-10-02 cs.CV 83%

Efficient Multi-modal Large Language Models via Progressive Consistency Distillation

Zichen Wen, Shaobo Wang, Yufa Zhou, Junyuan Zhang, Qintong Zhang, Yifeng Gao, Zhaorun Chen, Bin Wang, Weijia Li, Conghui He, Linfeng Zhang

专题命中效率与部署：large language model(title)；language model(title)

Comments Accepted by NeurIPS 2025

2510.00446 2025-10-02 cs.CL cs.SE 83%

LongCodeZip: Compress Long Context for Code Language Models

Yuling Shi, Yichun Qian, Hongyu Zhang, Beijun Shen, Xiaodong Gu

机构 * Shanghai Jiao Tong University（上海交通大学）； Stanford University（斯坦福大学）； Chongqing University（重庆大学）

专题命中效率与部署：language model(title,abstract)；large language model(abstract)；分类 cs.CL

Comments Accepted to ASE 2025. Code available at https://github.com/YerbaPage/LongCodeZip

2506.00910 2025-10-02 cs.LG cs.AI 81%

PCoreSet: Effective Active Learning through Knowledge Distillation from Vision-Language Models

Seongjae Kang, Dong Bok Lee, Hyungjoon Jang, Dongseop Kim, Sung Ju Hwang

机构 * VUNO Inc.（VUNO公司）； KAIST（韩国科学技术院）

专题命中效率与部署：language model(title,abstract)；分类 cs.AI、cs.LG

Comments 39 pages, 25 figures, preprint

2510.00841 2025-10-02 cs.LG 79%

LLM Routing with Dueling Feedback

Chao-Kai Chiang, Takashi Ishida, Masashi Sugiyama

机构 * The University of Tokyo（东京大学）； RIKEN AIP（日本理化学研究所AIP）

专题命中效率与部署：LLM(title,abstract)；分类 cs.LG

2509.25180 2025-10-02 cs.CV cs.AI 79%

DC-Gen: Post-Training Diffusion Acceleration with Deeply Compressed Latent Space

Wenkun He, Yuchao Gu, Junyu Chen, Dongyun Zou, Yujun Lin, Zhekai Zhang, Haocheng Xi, Muyang Li, Ligeng Zhu, Jincheng Yu, Junsong Chen, Enze Xie, Song Han, Han Cai

机构 * NVIDIA

专题命中效率与部署：post-training(title,abstract)；分类 cs.AI

Comments Tech Report. The first three authors contributed equally to this work

2507.14520 2025-10-02 cs.AI 79%

What if Othello-Playing Language Models Could See?

Xinyi Chen, Yifei Yuan, Jiaang Li, Serge Belongie, Maarten de Rijke, Anders Søgaard

机构 * University of Amsterdam（阿姆斯特丹大学）； ETH Zürich（苏黎世联邦理工学院）； University of Copenhagen（哥本哈根大学）

专题命中效率与部署：language model(title,abstract)；分类 cs.AI

Comments ICML 2025 Assessing World Models Workshop; EMNLP 2025 Findings

2502.01289 2025-10-02 cs.LG cs.CR cs.CV cs.DC 79%

A Framework for Double-Blind Federated Adaptation of Foundation Models

Nurbek Tastan, Karthik Nandakumar

机构 * Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)（穆罕默德·本·扎耶德人工智能大学）； Michigan State University (MSU)（密歇根州立大学）

专题命中效率与部署：foundation model(title,abstract)；分类 cs.LG

Comments Accepted to ICCV 2025

2510.00636 2025-10-02 cs.AI cs.CL 79%

Expected Attention: KV Cache Compression by Estimating Attention from Future Queries Distribution

Alessio Devoto, Maximilian Jeblick, Simon Jégou

机构 * Sapienza University of Rome（萨皮恩扎罗马大学）； NVIDIA（英伟达）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2510.00451 2025-10-02 cs.CR cs.AI cs.LG cs.MA 79%

A Call to Action for a Secure-by-Design Generative AI Paradigm

Dalal Alharthi, Ivan Roberto Kawaminami Garcia

机构 * University of Arizona（亚利桑那大学）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2510.01126 2025-10-02 cs.CV cs.RO 78%

Strategic Fusion of Vision Language Models: Shapley-Credited Context-Aware Dawid-Skene for Multi-Label Tasks in Autonomous Driving

Yuxiang Feng, Keyang Zhang, Hassane Ouchouid, Ashwil Kaniamparambil, Ioannis Souflas, Panagiotis Angeloudis

机构 * Centre for Transport Engineering and Modelling, Department of Civil and Environmental Engineering, Imperial College London（交通工程与建模中心，土木与环境工程系，帝国理工学院伦敦分校）； ELM Europe（ELM欧洲）

专题命中效率与部署：language model(title,abstract)

Comments 8 pages

2510.00313 2025-10-02 eess.AS cs.SD 78%

Post-Training Quantization for Audio Diffusion Transformers

Tanmay Khandelwal, Magdalena Fuentes

机构 * Courant Institute of Mathematical Sciences, New York University, NY, USA（Courant数学科学研究所，纽约大学）； MARL, New York University, NY, USA（MARL，纽约大学）； IDM, New York University, NY, USA（IDM，纽约大学）

专题命中效率与部署：post-training(title,abstract)

Comments 5 pages, 4 figures, accepted at IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) 2025

2509.23384 2025-10-02 cs.DC 78%

A Predictive and Synergistic Two-Layer Scheduling Framework for LLM Serving

Yue Zhang, Yuansheng Chen, Xuan Mo, Alex Xi, Jialun Li, WeiGang Wu

专题命中效率与部署：LLM(title,abstract)

Comments Update the system name in the summary

2510.00392 2025-10-02 q-bio.GN cs.CV cs.LG 77%

A Deep Learning Pipeline for Epilepsy Genomic Analysis Using GPT-2 XL and NVIDIA H100

Muhammad Omer Latif, Hayat Ullah, Muhammad Ali Shafique, Zhihua Dong

机构 * Connetquot Central School District of Long Island, NY 11716 USA（纽约长岛康奈克特中央学区）； Florida Atlantic University（佛罗里达大学）； Kansas State University（堪萨斯州立大学）； Brookhaven National Laboratory（布鲁赫斯国家实验室）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.LG

Comments 12 pages

2510.00452 2025-10-02 cs.CR cs.AI cs.LG cs.MA 73%

Cloud Investigation Automation Framework (CIAF): An AI-Driven Approach to Cloud Forensics

Dalal Alharthi, Ivan Roberto Kawaminami Garcia

机构 * University of Arizona（亚利桑那大学）

专题命中效率与部署：large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2510.00482 2025-10-02 cs.CL 70%

Agent Fine-tuning through Distillation for Domain-specific LLMs in Microdomains

Yawen Xue, Masaya Tsunokake, Yuta Koreeda, Ekant Muljibhai Amin, Takashi Sumiyoshi, Yasuhiro Sogawa

机构 * Research（研究）； Development Group Hitachi, Ltd. Tokyo, Japan（开发组日立株式会社东京日本）

专题命中效率与部署：large language model(abstract)；language model(abstract)；分类 cs.CL

Comments Accepted by AIxB 2025

2510.00067 2025-10-02 cs.CV cs.AI cs.HC 70%

Intelligent 5S Audit: Application of Artificial Intelligence for Continuous Improvement in the Automotive Industry

Rafael da Silva Maciel, Lucio Veraldo

机构 * Institute of Science and Technology（科学与技术研究所）； Federal University of São Paulo（圣保罗联邦大学）

专题命中效率与部署：LLM(abstract)；language model(abstract)；分类 cs.AI

Comments 8 pages, 5 figures, 5 tables

2510.01178 2025-10-02 cs.LG cs.AI 62%

COM-BOM: Bayesian Exemplar Search for Efficiently Exploring the Accuracy-Calibration Pareto Frontier

Gaoxiang Luo, Aryan Deshwal

机构 * University of Minnesota（明尼苏达大学）

专题命中效率与部署：LLM(abstract)；分类 cs.AI、cs.LG

Comments Accepted by EMNLP 2025 Main, Code: https://github.com/GaoxiangLuo/COM-BOM

2510.00078 2025-10-02 cs.LG cs.AI cs.DC 62%

Adaptive and Resource-efficient Agentic AI Systems for Mobile and Embedded Devices: A Survey

Sicong Liu, Weiye Wu, Xiangrui Xu, Teng Li, Bowen Pang, Bin Guo, Zhiwen Yu

专题命中效率与部署：foundation model(abstract)；分类 cs.AI、cs.LG

2509.25179 2025-10-02 cs.CL cs.AI 62%

NAIPv2: Debiased Pairwise Learning for Efficient Paper Quality Estimation

Penghai Zhao, Jinyu Tian, Qinghua Xing, Xin Zhang, Zheng Li, Jianjun Qian, Ming-Ming Cheng, Xiang Li

机构 * VCIP, CS, Nankai University（南开大学计算机科学与技术学院）； NKIARI, Shenzhen Futian（深圳福田NKIARI）； PCALab, School of Computer Science and Engineering, Nanjing University of Science and Technology（南京理工大学计算机科学与工程学院）

专题命中效率与部署：LLM(abstract)；分类 cs.CL、cs.AI

Comments NAIPv2 complements our earlier work NAIPv1 (arXiv:2408.03934). Whereas NAIPv1 addressed citation count-based impact prediction, NAIPv2 estimates research quality using peer review data

2507.16577 2025-10-02 cs.LG cs.CL 62%

Scaling Linear Attention with Sparse State Expansion

Yuqi Pan, Yongqi An, Zheng Li, Yuhong Chou, Ruijie Zhu, Xiaohui Wang, Mingxuan Wang, Jinqiao Wang, Guoqi Li

机构 * Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； The Hong Kong Polytechnic University（香港理工大学）

专题命中效率与部署：language model(abstract)；分类 cs.CL、cs.LG

2412.12351 2025-10-02 cs.LG cs.CL 62%

Krony-PT: GPT2 compressed with Kronecker Products

Mohamed Ayoub Ben Ayad, Jelena Mitrovic, Michael Granitzer

机构 * Chair of Data Science University of Passau（数据科学系大学）

专题命中效率与部署：language model(abstract)；分类 cs.CL、cs.LG

2510.00536 2025-10-02 cs.CL 57%

GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal Awareness

Kung-Hsiang Huang, Haoyi Qiu, Yutong Dai, Caiming Xiong, Chien-Sheng Wu

机构 * Salesforce AI Research（Salesforce AI研究）

专题命中效率与部署：language model(abstract)；分类 cs.CL

2510.00079 2025-10-02 cs.IT cs.LG math.IT stat.ML 57%

Directed Information $γ$-covering: An Information-Theoretic Framework for Context Engineering

Hai Huang

机构 * Atlassian（Atlassian公司）

专题命中效率与部署：LLM(abstract)；分类 cs.LG

Comments 15 pages, 6 tables, preprint

2510.00541 2025-10-02 cs.DC 50%

Towards Efficient VM Placement: A Two-Stage ACO-PSO Approach for Green Cloud Infrastructure

Ali M. Baydoun, Ahmed S. Zekri

专题命中效率与部署：prompting(abstract)

Comments 20 pages, 7 figures. Published in International Journal of Computer Networks & Communications (IJCNC), Vol. 17, No. 5, 2025

Journal ref Int. J. Comput. Netw. Commun. (IJCNC), Vol. 17, No. 5, pp. 117-136, 2025

2506.08678 2025-10-02 cs.CV 50%

ATAS: Any-to-Any Self-Distillation for Enhanced Open-Vocabulary Dense Prediction

Juan Yeo, Soonwoo Cha, Jiwoo Song, Hyunbin Jin, Taesup Kim

专题命中效率与部署：language model(abstract)

Comments Accepted at ICCV25

2505.13000 2025-10-02 cs.SD eess.AS 50%

DualCodec: A Low-Frame-Rate, Semantically-Enhanced Neural Audio Codec for Speech Generation

Jiaqi Li, Xiaolong Lin, Zhekai Li, Shixi Huang, Yuancheng Wang, Chaoren Wang, Zhenpeng Zhan, Zhizheng Wu

机构 * School of Data Science（数据科学学院）

专题命中效率与部署：language model(abstract)

Comments Accepted to Interspeech 2025

1. 效率与部署 30 篇

Combining Large Language Models and Gradient-Free Optimization for Automatic Control Policy Synthesis

Has the Two-Decade-Old Prophecy Come True? Artificial Bad Intelligence Triggered by Merely a Single-Bit Flip in Large Language Models

Rethinking RoPE Scaling in Quantized LLM: Theory, Outlier, and Channel-Band Analysis with Weight Rescaling

PromptPilot: Improving Human-AI Collaboration Through LLM-Enhanced Prompt Engineering

Efficient Multi-modal Large Language Models via Progressive Consistency Distillation

LongCodeZip: Compress Long Context for Code Language Models

PCoreSet: Effective Active Learning through Knowledge Distillation from Vision-Language Models

LLM Routing with Dueling Feedback

DC-Gen: Post-Training Diffusion Acceleration with Deeply Compressed Latent Space

What if Othello-Playing Language Models Could See?

A Framework for Double-Blind Federated Adaptation of Foundation Models

Expected Attention: KV Cache Compression by Estimating Attention from Future Queries Distribution

A Call to Action for a Secure-by-Design Generative AI Paradigm

Strategic Fusion of Vision Language Models: Shapley-Credited Context-Aware Dawid-Skene for Multi-Label Tasks in Autonomous Driving

Post-Training Quantization for Audio Diffusion Transformers

A Predictive and Synergistic Two-Layer Scheduling Framework for LLM Serving

A Deep Learning Pipeline for Epilepsy Genomic Analysis Using GPT-2 XL and NVIDIA H100

Cloud Investigation Automation Framework (CIAF): An AI-Driven Approach to Cloud Forensics

Agent Fine-tuning through Distillation for Domain-specific LLMs in Microdomains

Intelligent 5S Audit: Application of Artificial Intelligence for Continuous Improvement in the Automotive Industry

COM-BOM: Bayesian Exemplar Search for Efficiently Exploring the Accuracy-Calibration Pareto Frontier

Adaptive and Resource-efficient Agentic AI Systems for Mobile and Embedded Devices: A Survey

NAIPv2: Debiased Pairwise Learning for Efficient Paper Quality Estimation

Scaling Linear Attention with Sparse State Expansion

Krony-PT: GPT2 compressed with Kronecker Products

GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal Awareness

Directed Information $γ$-covering: An Information-Theoretic Framework for Context Engineering

Towards Efficient VM Placement: A Two-Stage ACO-PSO Approach for Green Cloud Infrastructure

ATAS: Any-to-Any Self-Distillation for Enhanced Open-Vocabulary Dense Prediction

DualCodec: A Low-Frame-Rate, Semantically-Enhanced Neural Audio Codec for Speech Generation