语言大模型 / LLM

2502.20220 2025-09-16 cs.CV 50%

Avat3r: Large Animatable Gaussian Reconstruction Model for High-fidelity 3D Head Avatars

Tobias Kirschstein, Javier Romero, Artem Sevastopolsky, Matthias Nießner, Shunsuke Saito

机构 * Technical University of Munich（慕尼黑技术大学）； Meta Reality Labs Pittsburgh（Meta现实实验室（匹兹堡））

专题命中评测与基准：foundation model(abstract)

Comments Project website: https://tobias-kirschstein.github.io/avat3r/, Video: https://youtu.be/P3zNVx15gYs

2504.01020 2025-09-16 cs.CV 50%

Shot-by-Shot: Film-Grammar-Aware Training-Free Audio Description Generation

Junyu Xie, Tengda Han, Max Bain, Arsha Nagrani, Eshika Khandelwal, Gül Varol, Weidi Xie, Andrew Zisserman

机构 * Visual Geometry Group, University of Oxford（牛津大学视觉几何组）； CVIT, IIIT Hyderabad（海得拉巴印度理工学院计算机视觉研究所）； LIGM, École des Ponts ParisTech（巴黎理工学院路易-狄塞尔数学与计算机科学实验室）； SAI, Shanghai Jiao Tong University（上海交通大学人工智能研究所）

专题命中评测与基准：language model(abstract)

Comments ICCV 2025. Project Page: https://www.robots.ox.ac.uk/vgg/research/shot-by-shot/

2406.17254 2025-09-16 cs.CV 50%

Scalp Diagnostic System With Label-Free Segmentation and Training-Free Image Translation

Youngmin Kim, Saejin Kim, Hoyeon Moon, Youngjae Yu, Junhyug Noh

机构 * Yonsei University（延世大学）； Seoul National University（首尔国立大学）； Ewha Womans University（成均馆大学）

专题命中评测与基准：prompting(abstract)

Comments Accepted to MICCAI 2025(https://papers.miccai.org/miccai-2025/0806-Paper5080.html), Project page: https://0110tpwls.github.io/scalpvision25/

2509.12019 2025-09-16 cs.LG cs.AI cs.CL 89%

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models

Sangjun Lee, Seung-taek Woo, Jungyu Jin, Changhun Lee, Eunhyeok Park

机构 * Graduate School of Artificial Intelligence（人工智能研究生院）； Department of Convergence IT Engineering（融合IT工程系）； Pohang University of Science and Technology (POSTECH)（釜山科学技术大学）

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

Comments EMNLP 2025 Main Conference, Long Paper (Oral)

2509.11868 2025-09-16 cs.CL cs.AI cs.HC cs.RO 88%

Growing Perspectives: Modelling Embodied Perspective Taking and Inner Narrative Development Using Large Language Models

Sabrina Patania, Luca Annese, Anna Lambiase, Anita Pellegrini, Tom Foulsham, Azzurra Ruggeri, Silvia Rossi, Silvia Serino, Dimitri Ognibene

机构 * Department of Psychology, University of Milan-Bicocca（心理学系，米兰-比科卡大学）； Department of Psychology, University of Essex（心理学系，埃塞克斯大学）； School of Social Sciences and Technology, Technical University of Munich（社会科学与技术学院，慕尼黑技术大学）； Department of Electrical Engineering and Information Technology, University of Naples Federico II（电气工程与信息技术系，那不勒斯费德里科二世大学）

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

Comments Accepted at ICDL https://icdl2025.fel.cvut.cz/

2509.12112 2025-09-16 cs.CL 88%

CBP-Tuning: Efficient Local Customization for Black-box Large Language Models

Jiaxuan Zhao, Naibin Gu, Yuchen Feng, Xiyu Liu, Peng Fu, Zheng Lin, Weiping Wang

机构 * Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）； School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络安全学院）

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

2508.17610 2025-09-16 cs.CL 88%

Less Is More? Examining Fairness in Pruned Large Language Models for Summarising Opinions

Nannan Huang, Haytham M. Fayek, Xiuzhen Zhang

机构 * RMIT University（皇家墨尔本理工大学）

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

Comments Accepted to EMNLP 2025 Main Conference

2509.12130 2025-09-16 cs.CL 86%

XplaiNLP at CheckThat! 2025: Multilingual Subjectivity Detection with Finetuned Transformers and Prompt-Based Inference with Large Language Models

Ariana Sahitaj, Jiaao Li, Pia Wenzel Neves, Fedor Splitt, Premtim Sahitaj, Charlott Jakob, Veronika Solopova, Vera Schmitt

机构 * Quality and Usability Lab, Technische Universität Berlin, Berlin, Germany（柏林技术大学质量与可用性实验室）； German Research Center for Artificial Intelligence (DFKI), Berlin, Germany（德国人工智能研究中心（DFKI））

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

2509.10753 2025-09-16 cs.LG cs.AI 86%

HalluField: Detecting LLM Hallucinations via Field-Theoretic Modeling

Minh Vu, Brian K. Tran, Syed A. Shah, Geigh Zollicoffer, Nhat Hoang-Xuan, Manish Bhattarai

机构 * Center for Nonlinear Studies（非线性研究中心）； Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室）； T-1 Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室T-1部门）； Applied Mathematics University of Colorado Boulder（科罗拉多大学博尔德分校应用数学系）； T-4 Los Alamos National Laboratory（洛斯阿拉莫斯国家实验室T-4部门）

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

2509.12152 2025-09-16 cs.HC cs.AI 85%

Beyond PII: How Users Attempt to Estimate and Mitigate Implicit LLM Inference

Synthia Wang, Sai Teja Peddinti, Nina Taft, Nick Feamster

机构 * University of Chicago（芝加哥大学）； Google（谷歌）

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

2509.06261 2025-09-16 cs.DC cs.LG 85%

FineServe: Precision-Aware KV Slab and Two-Level Scheduling for Heterogeneous Precision LLM Serving

Kyungmin Bin, Seungbeom Choi, Jimyoung Son, Jieun Choi, Daseul Bae, Daehyeon Baek, Kihyo Moon, Minsung Jang, Hyojung Lee

机构 * Cloud Research Team, Samsung SDS（三星SDS云研究团队）

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

2509.11524 2025-09-16 cs.IR 85%

Decoding in Latent Spaces for Efficient Inference in LLM-based Recommendation

Chengbing Wang, Yang Zhang, Zhicheng Wang, Tianhao Shi, Keqin Bao, Fuli Feng, Tat-Seng Chua

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

Comments Accepted for publication in EMNLP'25

2509.06516 2025-09-16 cs.LG cs.AI 81%

QualityFM: a Multimodal Physiological Signal Foundation Model with Self-Distillation for Signal Quality Challenges in Critically Ill Patients

Zongheng Guo, Tao Chen, Manuela Ferrario

机构 * Department of Electronics, Information and Bioengineering, Politecnico di Milano（电子、信息与生物工程系，米兰理工学院）； State Key Laboratory of Industrial Control Technology, Zhejiang University（工业控制技术国家重点实验室，浙江大学）

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

Comments 11 pages, 5 figures, 7 tables

2509.11155 2025-09-16 cs.LG cs.AI cs.CL 80%

AQUA: Attention via QUery mAgnitudes for Memory and Compute Efficient Inference in LLMs

Santhosh G S, Saurav Prakash, Balaraman Ravindran

机构 * Centre for Responsible AI（负责任人工智能中心）； Indian Institute of Technology Madras（印度理工学院马德拉斯分校）； Department of Electrical Engineering（电气工程系）； Wadhwani School of Data Science and Artificial Intelligence（Wadhwani数据科学与人工智能学院）

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

2509.10833 2025-09-16 cs.CL cs.AI cs.HC cs.LG 80%

Towards Automated Error Discovery: A Study in Conversational AI

Dominic Petrak, Thy Thy Tran, Iryna Gurevych

机构 * Ubiquitous Knowledge Processing Lab (UKP Lab), Department of Computer Science and Hessian Center for AI (hessian.AI), Technical University of Darmstadt（德累斯顿技术大学）

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

Comments Accepted to EMNLP 2025 main conference

2509.11961 2025-09-16 cs.CL 79%

Spec-LLaVA: Accelerating Vision-Language Models with Dynamic Tree-Based Speculative Decoding

Mingxiao Huo, Jiayi Zhang, Hewei Wang, Jinfeng Xu, Zheyu Chen, Huilin Tai, Yijun Chen

机构 * Carnegie Mellon University（卡内基梅隆大学）； University of Nottingham（诺丁汉大学）； The University of Hong Kong（香港大学）； The Hong Kong Polytechnic University（香港理工大学）； Columbia University（哥伦比亚大学）； University of California, Berkeley（加州大学伯克利分校）

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

Comments 7pages, accepted by ICML TTODLer-FM workshop

2504.05295 2025-09-16 cs.LG cs.AI math.OC 79%

Dion: Distributed Orthonormalized Updates

Kwangjun Ahn, Byron Xu, Natalie Abreu, Ying Fan, Gagik Magakyan, Pratyusha Sharma, Zheng Zhan, John Langford

机构 * Microsoft Research, AI Frontiers（微软研究院、人工智能前沿）； Harvard University（哈佛大学）； MIT（麻省理工学院）

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

Comments "Version 3" with various new updates

2504.12356 2025-09-16 eess.IV cs.CV 78%

Regist3R: Incremental Registration with Stereo Foundation Model

Sidun Liu, Wenyu Li, Peng Qiao, Yong Dou

机构 * College of Computer Science and Technology（计算机科学与技术学院）； National Key Laboratory of Parallel and Distributed Computing（并行与分布式计算国家重点实验室）； National University of Defense Technology（国防科技大学）

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

Comments Accepted by ACM Multimedia 2025. github link: https://github.com/Liu-SD/Regist3R

2504.10307 2025-09-16 cs.IR 78%

CROSSAN: Towards Efficient and Effective Adaptation of Multiple Multimodal Foundation Models for Sequential Recommendation

Junchen Fu, Yongxin Ni, Joemon M. Jose, Ioannis Arapakis, Kaiwen Zheng, Youhua Li, Xuri Ge

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

2509.10946 2025-09-16 cs.SE cs.AI 77%

When the Code Autopilot Breaks: Why LLMs Falter in Embedded Machine Learning

Roberto Morabito, Guanghan Wu

机构 * University of Helsinki（赫尔辛基大学）

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

Comments This paper has been accepted for publication in Computer (IEEE). Upon publication, the copyright will be transferred to IEEE

2509.08486 2025-09-16 cs.CL 77%

Too Helpful, Too Harmless, Too Honest or Just Right?

Gautam Siddharth Kashyap, Mark Dras, Usman Naseem

机构 * School of Computing, Macquarie University（计算机学院，麦考瑞大学）

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

Comments EMNLP'25 Main

2509.11628 2025-09-16 cs.LG cs.AI cs.CV 73%

SpeCa: Accelerating Diffusion Transformers with Speculative Feature Caching

Jiacheng Liu, Chang Zou, Yuanhuiyi Lyu, Fei Ren, Shaobo Wang, Kaixin Li, Linfeng Zhang

机构 * Shanghai Jiao Tong University（上海交通大学）； Shandong University（山东大学）； University of Electronic Science and Technology of China（电子科学与技术大学）； The Hong Kong University of Science and Technology (Guangzhou)（香港科学与技术大学（广州））； Tsinghua University（清华大学）； National University of Singapore（新加坡国立大学）

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

Comments 15 pages, 9 figures, ACM Multimedia 2025

2509.11374 2025-09-16 cs.CL cs.AI 73%

Transformer Enhanced Relation Classification: A Comparative Analysis of Contextuality, Data Efficiency and Sequence Complexity

Bowen Jing, Yang Cui, Tianpeng Huang

机构 * University of Manchester（曼彻斯特大学）

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

2509.11254 2025-09-16 math.OC cs.LG 70%

From PowerSGD to PowerSGD+: Low-Rank Gradient Compression for Distributed Optimization with Convergence Guarantees

Shengping Xie, Chuyan Chen, Kun Yuan

机构 * School of Mathematical Sciences, Peking University（北京大学数学科学学院）； Center for Machine Learning Research, Peking University（北京大学机器学习研究中心）

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

2509.10919 2025-09-16 cs.CV cs.LG 70%

Lightweight Metadata-Aware Mixture-of-Experts Masked Autoencoder for Earth Observation

Mohanad Albughdadi

机构 * European Centre for Medium-Range Weather Forecasts（欧洲中期天气预报中心）

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

2505.14864 2025-09-16 cs.DC cs.AI 70%

Balanced and Elastic End-to-end Training of Dynamic LLMs

Mohamed Wahib, Muhammed Abdullah Soyturk, Didem Unat

机构 * RIKEN Center for Computational Science（日本京都计算科学研究中心）； Koç University（科克大学）

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

2502.12455 2025-09-16 cs.CL 70%

DSMoE: Matrix-Partitioned Experts with Dynamic Routing for Computation-Efficient Dense LLMs

Minxuan Lv, Zhenpeng Su, Leiyu Pan, Yizhe Xiong, Zijia Lin, Hui Chen, Wei Zhou, Jungong Han, Guiguang Ding, Cheng Luo, Di Zhang, Kun Gai, Songlin Hu

机构 * Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）； University of Chinese Academy of Sciences（中国科学院大学）； Kuaishou Technology（快手科技）； Tianjin University（天津大学）； Tsinghua University（清华大学）

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

Comments Accepted by EMNLP main conference

2405.03228 2025-09-16 cs.LG 70%

TED: Accelerate Model Training by Internal Generalization

Jinying Xiao, Ping Li, Jie Nie

机构 * Changsha University of Science and Technology（长沙理工大学）

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

Comments ECAI 2024

2509.11134 2025-09-16 cs.DC 67%

GFS: A Preemption-aware Scheduling Framework for GPU Clusters with Predictive Spot Instance Management

Jiaang Duan, Shenglin Xu, Shiyou Qian, Dingyu Yang, Kangjin Wang, Chenzhi Liao, Yinghao Yu, Qin Hua, Hanwen Hu, Qi Wang, Wenchao Wu, Dongqing Bao, Tianyu Lu, Jian Cao, Guangtao Xue, Guodong Yang, Liping Zhang, Gang Chen

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

Comments This paper has been accepted to the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS 2026)

2509.00174 2025-09-16 cs.LG cs.AI 62%

Principled Approximation Methods for Efficient and Scalable Deep Learning

Pedro Savarese

机构 * TOYOTA TECHNOLOGICAL INSTITUTE AT CHICAGO（丰田技术研究所（芝加哥））

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

Comments PhD thesis

1. 评测与基准 62 篇

Avat3r: Large Animatable Gaussian Reconstruction Model for High-fidelity 3D Head Avatars

Shot-by-Shot: Film-Grammar-Aware Training-Free Audio Description Generation

Scalp Diagnostic System With Label-Free Segmentation and Training-Free Image Translation

2. 效率与部署 40 篇

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models

Growing Perspectives: Modelling Embodied Perspective Taking and Inner Narrative Development Using Large Language Models

CBP-Tuning: Efficient Local Customization for Black-box Large Language Models

Less Is More? Examining Fairness in Pruned Large Language Models for Summarising Opinions

XplaiNLP at CheckThat! 2025: Multilingual Subjectivity Detection with Finetuned Transformers and Prompt-Based Inference with Large Language Models

HalluField: Detecting LLM Hallucinations via Field-Theoretic Modeling

Beyond PII: How Users Attempt to Estimate and Mitigate Implicit LLM Inference

FineServe: Precision-Aware KV Slab and Two-Level Scheduling for Heterogeneous Precision LLM Serving

Decoding in Latent Spaces for Efficient Inference in LLM-based Recommendation

QualityFM: a Multimodal Physiological Signal Foundation Model with Self-Distillation for Signal Quality Challenges in Critically Ill Patients

AQUA: Attention via QUery mAgnitudes for Memory and Compute Efficient Inference in LLMs

Towards Automated Error Discovery: A Study in Conversational AI

Spec-LLaVA: Accelerating Vision-Language Models with Dynamic Tree-Based Speculative Decoding

Dion: Distributed Orthonormalized Updates

Regist3R: Incremental Registration with Stereo Foundation Model

CROSSAN: Towards Efficient and Effective Adaptation of Multiple Multimodal Foundation Models for Sequential Recommendation

When the Code Autopilot Breaks: Why LLMs Falter in Embedded Machine Learning

Too Helpful, Too Harmless, Too Honest or Just Right?

SpeCa: Accelerating Diffusion Transformers with Speculative Feature Caching

Transformer Enhanced Relation Classification: A Comparative Analysis of Contextuality, Data Efficiency and Sequence Complexity

From PowerSGD to PowerSGD+: Low-Rank Gradient Compression for Distributed Optimization with Convergence Guarantees

Lightweight Metadata-Aware Mixture-of-Experts Masked Autoencoder for Earth Observation

Balanced and Elastic End-to-end Training of Dynamic LLMs

DSMoE: Matrix-Partitioned Experts with Dynamic Routing for Computation-Efficient Dense LLMs

TED: Accelerate Model Training by Internal Generalization

GFS: A Preemption-aware Scheduling Framework for GPU Clusters with Predictive Spot Instance Management

Principled Approximation Methods for Efficient and Scalable Deep Learning