大模型推理能力 - arXivDaily 专题

2511.07299 2025-12-15 cs.CV 50%

VADER: Towards Causal Video Anomaly Understanding with Relation-Aware Large Language Models

VADER：基于关系感知大型语言模型的因果视频异常理解

Ying Cheng, Yu-Ho Lin, Min-Hung Chen, Fu-En Yang, Shang-Hong Lai

机构 * National Tsing Hua University（清华大学）； NVIDIA

专题命中测试时计算：reasoning(abstract)

AI总结 VADER通过整合关系感知的大型语言模型与视频关键帧特征，提升视频异常的因果理解与解释能力。

Comments Accepted to WACV 2026. Project page available at: https://vader-vau.github.io/

2512.08079 2025-12-10 cs.IR 50%

Leveraging Machine Learning and Large Language Models for Automated Image Clustering and Description in Legal Discovery

利用机器学习和大语言模型实现法律发现中的自动化图像聚类与描述

Qiang Mao, Fusheng Wei, Robert Neary, Charles Wang, Han Qin, Jianping Zhang, Nathaniel Huber-Fliflet

专题命中测试时计算：chain-of-thought(abstract)

AI总结本文提出利用机器学习和大语言模型实现法律发现中自动化图像聚类与描述，通过比较不同采样策略、提示技术和生成方法，提升大规模图像数据处理的效率和准确性。

2512.06017 2025-12-09 cs.RO eess.IV 50%

Training-Free Robot Pose Estimation using Off-the-Shelf Foundational Models

无需训练的机器人姿态估计使用现成的基础模型

Laurence Liang

机构 * McGill University（麦吉尔大学）

专题命中测试时计算：verifier(abstract)

AI总结本文提出利用现成的基础模型无需训练即可估计机器人关节角度，通过实验验证了模型性能基准及测试扩展对预测效果的影响。

Comments Accepted at CVIS 2025

2511.20390 2025-11-26 cs.CV 50%

FREE: Uncertainty-Aware Autoregression for Parallel Diffusion Transformers

FREE: 用于并行扩散变换器的不确定性感知自回归

Xinwan Wen, Bowen Li, Jiajun Luo, Ye Li, Zhi Wang

机构 * Tsinghua University（清华大学）

专题命中测试时计算：verifier(abstract)

AI总结 FREE通过引入不确定性感知的自回归方法，在并行扩散变换器中实现高效的无损加速，提升去噪效率的同时保持生成质量。

2511.16951 2025-11-24 cs.CV 50%

FingerCap: Fine-grained Finger-level Hand Motion Captioning

FingerCap：细粒度指尖级手部动作描述

Xin Shen, Rui Zhu, Lei Shen, Xinyu Wang, Kaihao Zhang, Tianqing Zhu, Shuchen Wu, Chenxi Miao, Weikang Li, Yang Li, Deguo Xia, Jizhou Huang, Xin Yu

机构 * Baidu Inc.（百度公司）； The University of Queensland（昆士兰大学）； Nanjing University（南京大学）； Institute of Computing Technology, CAS（中国科学院计算技术研究所）； Peking University（北京大学）； Australian National University（澳大利亚国立大学）； City University of Macau（澳门城市大学）

专题命中测试时计算：reasoning(abstract)

AI总结 FingerCap通过FiGOP技术提升细粒度手部动作描述的准确性，解决视频大语言模型在指尖运动识别中的不足。

2511.10648 2025-11-14 cs.CV 50%

Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling

Jiahao Wang, Weiye Xu, Aijun Yang, Wengang Zhou, Lewei Lu, Houqiang Li, Xiaohua Wang, Jinguo Zhu

机构 * Xi’an Jiaotong University（西安交通大学）； University of Science and Technology of China（中国科学技术大学）； SenseTime Research（商汤科技研究院）

专题命中测试时计算：reasoning(abstract)

Comments Accepted to NeurIPS 2025 (The Thirty-Ninth Annual Conference on Neural Information Processing Systems)

2508.17714 2025-11-11 cs.CV 50%

F2RVLM: Boosting Fine-grained Fragment Retrieval for Multi-Modal Long-form Dialogue with Vision Language Model

Hanbo Bi, Zhiqiang Yuan, Zexi Jia, Jiapei Zhang, Chongyang Li, Peixiang Luo, Ying Deng, Xiaoyue Duan, Jinchao Zhang

专题命中测试时计算：reasoning(abstract)

2511.01423 2025-11-04 cs.SE 50%

LLM-Assisted Tool for Joint Generation of Formulas and Functions in Rule-Based Verification of Map Transformations

Ruidi He, Yu Zhang, Meng Zhang, Andreas Rausch

专题命中测试时计算：verifier(abstract)

2511.01208 2025-11-04 cs.IR 50%

Contextual Relevance and Adaptive Sampling for LLM-Based Document Reranking

Jerry Huang, Siddarth Madala, Cheng Niu, Julia Hockenmaier, Tong Zhang

专题命中测试时计算：reasoning(abstract)

2510.14624 2025-10-17 cs.CV 50%

Efficient Video Sampling: Pruning Temporally Redundant Tokens for Faster VLM Inference

Natan Bagrov, Eugene Khvedchenia, Borys Tymchenko, Shay Aharon, Lior Kadoch, Tomer Keren, Ofri Masad, Yonatan Geifman, Ran Zilberstein, Tuomas Rintamaki, Matthieu Le, Andrew Tao

机构 * NVIDIA

专题命中测试时计算：reasoning(abstract)

2503.11070 2025-10-17 cs.CV 50%

Falcon: A Remote Sensing Vision-Language Foundation Model (Technical Report)

Kelu Yao, Nuo Xu, Rong Yang, Yingying Xu, Zhuoyan Gao, Titinunt Kitrungrotsakul, Yi Ren, Pu Zhang, Jin Wang, Ning Wei, Chao Li

机构 * Research Center for Space Computing System, ZhejiangLab, Hangzhou, China（空间计算系统研究中心，浙江实验室，杭州，中国）

专题命中测试时计算：reasoning(abstract)

2510.09274 2025-10-13 cs.CV 50%

MomentSeg: Moment-Centric Sampling for Enhanced Video Pixel Understanding

Ming Dai, Sen Yang, Boqiang Duan, Wankou Yang, Jingdong Wang

机构 * Southeast University（东南大学）； Baidu VIS（百度视觉）

专题命中测试时计算：reasoning(abstract)

2510.04041 2025-10-07 cs.RO 50%

SITCOM: Scaling Inference-Time COMpute for VLAs

Ayudh Saxena, Harsh Shah, Sandeep Routray, Rishi Rajesh Shah, Esha Pahwa

机构 * Carnegie Mellon University（卡内基梅隆大学）

专题命中测试时计算：planning(abstract)

Comments Accepted at the NeurIPS 2025 Workshop on Space in Vision, Language, and Embodied AI (SpaVLE). *Equal contribution

2503.04130 2025-09-24 cs.CV 50%

STORM: Token-Efficient Long Video Understanding for Multimodal LLMs

Jindong Jiang, Xiuyu Li, Zhijian Liu, Muyang Li, Guo Chen, Zhiqi Li, De-An Huang, Guilin Liu, Zhiding Yu, Kurt Keutzer, Sungjin Ahn, Jan Kautz, Hongxu Yin, Yao Lu, Song Han, Wonmin Byeon

机构 * NVIDIA（英伟达）； Rutgers University（罗格斯大学）； UC Berkeley（加州大学伯克利分校）； MIT（麻省理工学院）； Nanjing University（南京大学）； KAIST（韩国科学技术院）

专题命中测试时计算：reasoning(abstract)

2509.17522 2025-09-23 cs.CV 50%

Chat-CBM: Towards Interactive Concept Bottleneck Models with Frozen Large Language Models

Hangzhou He, Lei Zhu, Kaiwen Li, Xinliang Zhang, Jiakui Hu, Ourui Fu, Zhengjian Yao, Yanye Lu

机构 * Department of Biomedical Engineering, College of Future Technology, Peking University（生物医学工程系，未来技术学院，北京大学）； Institute of Medical Technology, Peking University Health Science Center, Peking University（医学技术研究所，北京大学医学部，北京大学）； National Biomedical Imaging Center, College of Future Technology, Peking University（国家生物医学成像中心，未来技术学院，北京大学）

专题命中测试时计算：reasoning(abstract)

2509.11238 2025-09-16 cs.SE 50%

UserTrace: User-Level Requirements Generation and Traceability Recovery from Software Project Repositories

Dongming Jin, Zhi Jin, Yiran Zhang, Zheng Fang, Linyu Li, Yuanpeng He, Xiaohong Chen, Weisong Sun

专题命中测试时计算：verifier(abstract)

Comments 21page, 5 figures

2509.09263 2025-09-12 cs.CV 50%

DATE: Dynamic Absolute Time Enhancement for Long Video Understanding

Chao Yuan, Yang Yang, Yehui Yang, Zach Cheng

机构 * Beihang University（北京航空航天大学）； Dcar, ByteDance（字节跳动Dcar部门）； Qfin Holdings,Inc（Qfin控股公司）； MAIS, Institute of Automation, Chinese Academy of Sciences（自动化研究所，中国科学院）

专题命中测试时计算：reasoning(abstract)

2503.19903 2025-08-05 cs.CV 50%

Scaling Vision Pre-Training to 4K Resolution

Baifeng Shi, Boyi Li, Han Cai, Yao Lu, Sifei Liu, Marco Pavone, Jan Kautz, Song Han, Trevor Darrell, Pavlo Molchanov, Hongxu Yin

机构 * UC Berkeley（加州大学伯克利分校）； NVIDIA（英伟达）

专题命中测试时计算：test-time compute(abstract)

Comments CVPR 2025. Project Page: https://nvlabs.github.io/PS3

2504.14641 2025-07-28 cs.SE cs.SY eess.SY 50%

HLSTester: Efficient Testing of Behavioral Discrepancies with LLMs for High-Level Synthesis

Kangwei Xu, Bing Li, Grace Li Zhang, Ulf Schlichtmann

专题命中测试时计算：reasoning(abstract)

Comments arXiv admin note: text overlap with arXiv:2407.03889

2506.04501 2025-06-06 cs.CV 50%

AuthGuard: Generalizable Deepfake Detection via Language Guidance

Guangyu Shen, Zhihua Li, Xiang Xu, Tianchen Zhao, Zheng Zhang, Dongsheng An, Zhuowen Tu, Yifan Xing, Qin Zhang

机构 * Purdue University（普渡大学）； AWS AI Labs（AWS人工智能实验室）

专题命中测试时计算：reasoning(abstract)

2410.04249 2025-05-07 cs.SE 50%

DiffSpec: Differential Testing with LLMs using Natural Language Specifications and Code Artifacts

Nikitha Rao, Elizabeth Gilbert, Harrison Green, Tahina Ramananandro, Nikhil Swamy, Claire Le Goues, Sarah Fakhoury

专题命中测试时计算：verifier(abstract)

2504.17525 2025-04-25 cs.CV 50%

Text-to-Image Alignment in Denoising-Based Models through Step Selection

Paul Grimal, Hervé Le Borgne, Olivier Ferret

机构 * Université Paris-Saclay, CEA, List（巴黎-萨克雷大学、法国原子能委员会、List）

专题命中测试时计算：reasoning(abstract)

2504.06835 2025-04-10 cs.CV 50%

LVC: A Lightweight Compression Framework for Enhancing VLMs in Long Video Understanding

Ziyi Wang, Haoran Wu, Yiming Rong, Deyang Jiang, Yixin Zhang, Yunlong Zhao, Shuang Xu, Bo XU

专题命中测试时计算：reasoning(abstract)

2503.12496 2025-03-31 cs.CV 50%

Does Your Vision-Language Model Get Lost in the Long Video Sampling Dilemma?

Tianyuan Qu, Longxiang Tang, Bohao Peng, Senqiao Yang, Bei Yu, Jiaya Jia

专题命中测试时计算：reasoning(abstract)

2503.12271 2025-03-18 cs.CV 50%

Reflect-DiT: Inference-Time Scaling for Text-to-Image Diffusion Transformers via In-Context Reflection

Shufan Li, Konstantinos Kallidromitis, Akash Gokul, Arsh Koneru, Yusuke Kato, Kazuki Kozuka, Aditya Grover

专题命中测试时计算：reasoning(abstract)

Comments 17 pages, 9 figures

2502.06428 2025-02-12 cs.CV 50%

CoS: Chain-of-Shot Prompting for Long Video Understanding

Jian Hu, Zixu Cheng, Chenyang Si, Wei Li, Shaogang Gong

专题命中测试时计算：reasoning(abstract)

Comments A training-free test-time optimisation approach for long video understanding

2501.16125 2025-02-12 cs.IR 50%

SampleLLM: Optimizing Tabular Data Synthesis in Recommendations

Jingtong Gao, Zhaocheng Du, Xiaopeng Li, Yichao Wang, Xiangyang Li, Huifeng Guo, Ruiming Tang, Xiangyu Zhao

专题命中测试时计算：chain-of-thought(abstract)

2502.00028 2025-02-04 cs.AR cs.PL 50%

VRank: Enhancing Verilog Code Generation from Large Language Models via Self-Consistency

Zhuorui Zhao, Ruidi Qiu, Ing-Chao Lin, Grace Li Zhang, Bing Li, Ulf Schlichtmann

专题命中测试时计算：chain-of-thought(abstract)

Comments accepted by ISQED2025

2407.21534 2025-01-08 cs.CV 50%

ControlMLLM: Training-Free Visual Prompt Learning for Multimodal Large Language Models

Mingrui Wu, Xinyue Cai, Jiayi Ji, Jiale Li, Oucheng Huang, Gen Luo, Hao Fei, Guannan Jiang, Xiaoshuai Sun, Rongrong Ji

专题命中测试时计算：reasoning(abstract)

Comments Accepted to NeurIPS 2024; Code:https://github.com/mrwu-mac/ControlMLLM

2501.01476 2025-01-06 physics.data-an 50%

Sparse identification of evolution equations via Bayesian model selection

Tim W. Kroll, Oliver Kamps

专题命中测试时计算：reasoning(abstract)

Comments 22 pages, 6 fgures

1. 测试时计算 2047 篇

VADER: Towards Causal Video Anomaly Understanding with Relation-Aware Large Language Models

Leveraging Machine Learning and Large Language Models for Automated Image Clustering and Description in Legal Discovery

Training-Free Robot Pose Estimation using Off-the-Shelf Foundational Models

FREE: Uncertainty-Aware Autoregression for Parallel Diffusion Transformers

FingerCap: Fine-grained Finger-level Hand Motion Captioning

Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling

F2RVLM: Boosting Fine-grained Fragment Retrieval for Multi-Modal Long-form Dialogue with Vision Language Model

LLM-Assisted Tool for Joint Generation of Formulas and Functions in Rule-Based Verification of Map Transformations

Contextual Relevance and Adaptive Sampling for LLM-Based Document Reranking

Efficient Video Sampling: Pruning Temporally Redundant Tokens for Faster VLM Inference

Falcon: A Remote Sensing Vision-Language Foundation Model (Technical Report)

MomentSeg: Moment-Centric Sampling for Enhanced Video Pixel Understanding

SITCOM: Scaling Inference-Time COMpute for VLAs

STORM: Token-Efficient Long Video Understanding for Multimodal LLMs

Chat-CBM: Towards Interactive Concept Bottleneck Models with Frozen Large Language Models

UserTrace: User-Level Requirements Generation and Traceability Recovery from Software Project Repositories

DATE: Dynamic Absolute Time Enhancement for Long Video Understanding

Scaling Vision Pre-Training to 4K Resolution

HLSTester: Efficient Testing of Behavioral Discrepancies with LLMs for High-Level Synthesis

AuthGuard: Generalizable Deepfake Detection via Language Guidance

DiffSpec: Differential Testing with LLMs using Natural Language Specifications and Code Artifacts

Text-to-Image Alignment in Denoising-Based Models through Step Selection

LVC: A Lightweight Compression Framework for Enhancing VLMs in Long Video Understanding

Does Your Vision-Language Model Get Lost in the Long Video Sampling Dilemma?

Reflect-DiT: Inference-Time Scaling for Text-to-Image Diffusion Transformers via In-Context Reflection

CoS: Chain-of-Shot Prompting for Long Video Understanding

SampleLLM: Optimizing Tabular Data Synthesis in Recommendations

VRank: Enhancing Verilog Code Generation from Large Language Models via Self-Consistency

ControlMLLM: Training-Free Visual Prompt Learning for Multimodal Large Language Models

Sparse identification of evolution equations via Bayesian model selection