语言大模型 / LLM - arXivDaily 专题

2507.20169 2025-07-29 cs.SD eess.AS 86%

Self-Improvement for Audio Large Language Model using Unlabeled Speech

Shaowen Wang, Xinyuan Chen, Yao Xu

机构 * University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）

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

Comments To appear in Interspeech 2025. 6 pages, 1 figure

2505.22447 2025-05-29 cs.CR 86%

Privacy-preserving Prompt Personalization in Federated Learning for Multimodal Large Language Models

Sizai Hou, Songze Li, Baturalp Buyukates

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

Comments Under Review

2505.07203 2025-05-13 cs.DC 86%

PrefillOnly: An Inference Engine for Prefill-only Workloads in Large Language Model Applications

Kuntai Du, Bowen Wang, Chen Zhang, Yiming Cheng, Qing Lan, Hejian Sang, Yihua Cheng, Jiayi Yao, Xiaoxuan Liu, Yifan Qiao, Ion Stoica, Junchen Jiang

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

2403.09583 2025-04-18 cs.RO 86%

ExploRLLM: Guiding Exploration in Reinforcement Learning with Large Language Models

Runyu Ma, Jelle Luijkx, Zlatan Ajanovic, Jens Kober

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

Comments 6 pages, 6 figures, IEEE International Conference on Robotics and Automation (ICRA) 2025

2504.07787 2025-04-11 cs.SE 86%

Fairness Mediator: Neutralize Stereotype Associations to Mitigate Bias in Large Language Models

Yisong Xiao, Aishan Liu, Siyuan Liang, Xianglong Liu, Dacheng Tao

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

Comments Accepted by ISSTA 2025.20 pages

2503.18565 2025-03-25 cs.LG cs.AI cs.CL 86%

Distil-xLSTM: Learning Attention Mechanisms through Recurrent Structures

Abdoul Majid O. Thiombiano, Brahim Hnich, Ali Ben Mrad, Mohamed Wiem Mkaouer

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

2503.03154 2025-03-07 cs.HC 86%

Dango: A Mixed-Initiative Data Wrangling System using Large Language Model

Wei-Hao Chen, Weixi Tong, Amanda Case, Tianyi Zhang

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

Comments To appear in the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25), April 26-May 1, 2025, Yokohama, Japan

2501.02850 2025-01-07 cs.CV 86%

Large Language Models for Video Surveillance Applications

Ulindu De Silva, Leon Fernando, Billy Lau Pik Lik, Zann Koh, Sam Conrad Joyce, Belinda Yuen, Chau Yuen

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

Comments Accepted for TENCON 2024

2412.20820 2024-12-31 eess.SP cs.ET 86%

Retrieval-Augmented Generation for Mobile Edge Computing via Large Language Model

Runtao Ren, Yinyu Wu, Xuhui Zhang, Jinke Ren, Yanyan Shen, Shuqiang Wang, Kim-Fung Tsang

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

Comments This manuscript has been submitted to IEEE

2412.20322 2024-12-31 cs.AR cs.DC 86%

GreenLLM: Disaggregating Large Language Model Serving on Heterogeneous GPUs for Lower Carbon Emissions

Tianyao Shi, Yanran Wu, Sihang Liu, Yi Ding

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

Comments 15 pages, 15 figures

2411.11933 2024-12-02 cs.LG cs.AI cs.CL 86%

METEOR: Evolutionary Journey of Large Language Models from Guidance to Self-Growth

Jiawei Li, Xiaoang Xu, Yang Gao

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

Comments Our code can be found at https://github.com/DIRECT-BIT/METEOR

2411.12448 2024-11-25 cs.CV eess.IV 86%

Large Language Models for Lossless Image Compression: Next-Pixel Prediction in Language Space is All You Need

Kecheng Chen, Pingping Zhang, Hui Liu, Jie Liu, Yibing Liu, Jiaxin Huang, Shiqi Wang, Hong Yan, Haoliang Li

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

2410.17370 2024-10-24 cs.SE 86%

On the use of Large Language Models in Model-Driven Engineering

Juri Di Rocco, Davide Di Ruscio, Claudio Di Sipio, Phuong T. Nguyen, Riccardo Rubei

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

Comments A paper submitted to the Software Systems and Modeling Journal (Springer), and it has undergone the second revision

2409.16019 2024-09-25 cs.RO 86%

AIR-Embodied: An Efficient Active 3DGS-based Interaction and Reconstruction Framework with Embodied Large Language Model

Zhenghao Qi, Shenghai Yuan, Fen Liu, Haozhi Cao, Tianchen Deng, Jianfei Yang, Lihua Xie

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

2406.06385 2024-09-04 cs.LG cs.AI cs.CL 86%

Low-Rank Quantization-Aware Training for LLMs

Yelysei Bondarenko, Riccardo Del Chiaro, Markus Nagel

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

2407.12723 2024-07-18 cs.HC cs.CY 86%

The Future of Learning: Large Language Models through the Lens of Students

He Zhang, Jingyi Xie, Chuhao Wu, Jie Cai, ChanMin Kim, John M. Carroll

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

2406.00030 2024-06-04 cs.CL cs.AI cs.LG 86%

Large Language Model Pruning

Hanjuan Huang, Hao-Jia Song, Hsing-Kuo Pao

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

Comments 17 pages, 7 figures, 2 tables

2405.20985 2024-06-03 cs.CV 86%

DeCo: Decoupling Token Compression from Semantic Abstraction in Multimodal Large Language Models

Linli Yao, Lei Li, Shuhuai Ren, Lean Wang, Yuanxin Liu, Xu Sun, Lu Hou

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

2405.16009 2024-05-28 cs.CV 86%

Streaming Long Video Understanding with Large Language Models

Rui Qian, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Shuangrui Ding, Dahua Lin, Jiaqi Wang

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

2210.03696 2024-05-28 cs.CL cs.AI cs.LG cs.SE 86%

LLMEffiChecker: Understanding and Testing Efficiency Degradation of Large Language Models

Xiaoning Feng, Xiaohong Han, Simin Chen, Wei Yang

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

Comments This paper has been accepted to the TOSEM 2024

2402.09543 2024-02-16 cs.IR 86%

Rethinking Large Language Model Architectures for Sequential Recommendations

Hanbing Wang, Xiaorui Liu, Wenqi Fan, Xiangyu Zhao, Venkataramana Kini, Devendra Yadav, Fei Wang, Zhen Wen, Jiliang Tang, Hui Liu

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

Comments 8 pages, 5 figures, conference

2310.04573 2023-10-10 cs.LG cs.AI cs.CL 86%

Can pruning make Large Language Models more efficient?

Sia Gholami, Marwan Omar

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

2310.02421 2023-10-05 cs.LG cs.AI cs.CL 86%

Can a student Large Language Model perform as well as it's teacher?

Sia Gholami, Marwan Omar

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

2608.19147 2026-08-20 cs.DC cs.AI cs.SE 新提交 85%

Pre-Compiled Pipeline Shards for Distributed LLM Inference on Intel AI PC Fleets

面向英特尔AI PC集群的分布式大语言模型推理的预编译流水线分片

Tate Berenbaum, Muthaiah Venkatachalam

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

AI总结本文提出基于预编译流水线分片的分布式LLM推理方案，借助OpenVINO优化、投机解码与微批处理，实现AI PC集群高效运行单设备无法承载的大模型，提升了吞吐量与交互速度。

2502.00350 2026-08-20 cs.SE cs.AI 85%

OrcaLoca: An LLM Agent Framework for Software Issue Localization

Zhongming Yu, Hejia Zhang, Yujie Zhao, Hanxian Huang, Matrix Yao, Ke Ding, Jishen Zhao

机构 * University of California, San Diego, USA（加州大学圣迭戈分校）； Intel Corporation（英特尔公司）

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

2608.15567 2026-08-18 cs.LG 新提交 85%

SchurQuant: Groupwise Discrete Optimization for Layer-Wise LLM Quantization

SchurQuant：面向层大型语言模型量化的分组离散优化

Gunjun Lee, Sehwan Son, Younjoo Lee, Byungjun Kim, Jung Ho Ahn

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

AI总结 SchurQuant结合SCHUROPT与多种技术，在8个Llama和Qwen模型的2位量化任务中，使无反向传播PTQ的平均零样本精度比最强基线高9.65pp，提升了2位大语言模型量化的精度。

Comments 14 pages, 6 tables

2608.15282 2026-08-18 cs.LG cs.CV physics.bio-ph 新提交 85%

Earth Observation Foundation Models for Terrestrial Ecohydrology: From Representation Learning to Process Inference

面向陆地生态水文学的地球观测基础模型：从表示学习到过程推理

Yi Yu, Jian Peng, Yucheng Lin, Trevor F. Keenan, Thomas F. A. Bishop

机构 * The University of Sydney（悉尼大学）； Helmholtz Centre for Environmental Research–UFZ（亥姆霍兹环境研究中心）； Leipzig University（莱比锡大学）； City University of Hong Kong（香港城市大学）； University of California, Berkeley（加州大学伯克利分校）； Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室）

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

AI总结该研究构建框架分析EOFM在生态水文学中的应用，发现其与生态水文学需求存在不匹配，推动建立过程感知框架以支撑对水、能量与碳耦合动态的可信监测与解释。

2607.07964 2026-08-11 cs.LG 版本更新 85%

KronQ: LLM Quantization via Kronecker-Factored Hessian

KronQ：通过克罗内克分解海森矩阵进行大语言模型量化

Donghyun Lee, Yuhang Li, Ruokai Yin, Priyadarshini Panda

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

AI总结研究提出KronQ框架，通过引入梯度协方差挑战现有二阶PTQ方法假设。利用克罗内克分解海森矩阵近似，在两个层面改进量化：双向非相干处理减少权重幅度方差，导出新敏感度度量用于层间混合精度分配，在特定量化任务中表现优于GPTQ等。

Comments COLM 2026

2608.06901 2026-08-10 cs.CV cs.LG 新提交 85%

Prune Once: Retraining-Free Task-Agnostic Pruning for Vision-Language Models

一次剪枝：面向视觉-语言模型的无需重训练的任务无关剪枝

Minseok Kang, Hyunwoo Kim, Chanyoung Kim, Minwoo Kim, Jaekoo Lee, Dahuin Jung

机构 * Chung-Ang University（中央大学）； Soongsil University（崇实大学）； Kookmin University（国民大学）

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

AI总结本文提出名为PORTA的无需重训练的任务无关VLM剪枝框架，通过自适应稀疏分配实现高压缩下的竞争力性能，支持高效VLM压缩。

Comments Accepted to ECCV 2026

2607.29678 2026-08-10 cs.CL cs.DC cs.PF 版本更新 85%

TokTier: Exact Stateful CPU+GPU Tokenization for Agentic LLM Serving

TokTier：面向智能体大语言模型服务的精确有状态分词方案

Zhenyu Zhang, Zhichao Cao

机构 * Arizona State University（亚利桑那州立大学）

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

AI总结本研究提出TokTier，一种面向智能体LLM服务的精确有状态分词方案，通过增量修复、GPU加速分词等技术，将首次令牌生成时间降低16%-34%，饱和请求速率提升至1821次/秒，性能远超现有方案。

Comments 26 pages. Code: https://github.com/asu-idi/toktier

1. 效率与部署 22313 篇

Self-Improvement for Audio Large Language Model using Unlabeled Speech

Privacy-preserving Prompt Personalization in Federated Learning for Multimodal Large Language Models

PrefillOnly: An Inference Engine for Prefill-only Workloads in Large Language Model Applications

ExploRLLM: Guiding Exploration in Reinforcement Learning with Large Language Models

Fairness Mediator: Neutralize Stereotype Associations to Mitigate Bias in Large Language Models

Distil-xLSTM: Learning Attention Mechanisms through Recurrent Structures

Dango: A Mixed-Initiative Data Wrangling System using Large Language Model

Large Language Models for Video Surveillance Applications

Retrieval-Augmented Generation for Mobile Edge Computing via Large Language Model

GreenLLM: Disaggregating Large Language Model Serving on Heterogeneous GPUs for Lower Carbon Emissions

METEOR: Evolutionary Journey of Large Language Models from Guidance to Self-Growth

Large Language Models for Lossless Image Compression: Next-Pixel Prediction in Language Space is All You Need

On the use of Large Language Models in Model-Driven Engineering

AIR-Embodied: An Efficient Active 3DGS-based Interaction and Reconstruction Framework with Embodied Large Language Model

Low-Rank Quantization-Aware Training for LLMs

The Future of Learning: Large Language Models through the Lens of Students

Large Language Model Pruning

DeCo: Decoupling Token Compression from Semantic Abstraction in Multimodal Large Language Models

Streaming Long Video Understanding with Large Language Models

LLMEffiChecker: Understanding and Testing Efficiency Degradation of Large Language Models

Rethinking Large Language Model Architectures for Sequential Recommendations

Can pruning make Large Language Models more efficient?

Can a student Large Language Model perform as well as it's teacher?

Pre-Compiled Pipeline Shards for Distributed LLM Inference on Intel AI PC Fleets

OrcaLoca: An LLM Agent Framework for Software Issue Localization

SchurQuant: Groupwise Discrete Optimization for Layer-Wise LLM Quantization

Earth Observation Foundation Models for Terrestrial Ecohydrology: From Representation Learning to Process Inference

KronQ: LLM Quantization via Kronecker-Factored Hessian

Prune Once: Retraining-Free Task-Agnostic Pruning for Vision-Language Models

TokTier: Exact Stateful CPU+GPU Tokenization for Agentic LLM Serving