arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

Imperial College London(帝国理工学院)

共收录 97
2601.22818 2026-07-20 cs.CR cs.AI 版本更新

Hide and Seek in Embedding Space: Geometry-based Steganography and Detection in Large Language Models

嵌入空间中的隐秘行动:基于几何的隐写术与大语言模型中的检测

Charles Westphal, Keivan Navaie, Fernando E. Rosas

机构 * UCL Centre for Artificial Intelligence, University College London, UK(伦敦大学学院人工智能中心,大学学院伦敦) School of Computing and Communications, Lancaster University, UK(兰卡斯特大学计算机与通讯学院) Department of Informatics, University of Sussex, UK(苏塞克斯大学信息学院) Centre for Psychedelic Research and Centre for Complexity Science, Imperial College London, UK(伦敦帝国学院迷幻研究与复杂科学中心) Centre for Eudaimonia and Human Flourishing, University of Oxford, UK(牛津大学幸福与人类繁荣中心)

AI总结 本研究提出了一种低恢复性隐写术,通过嵌入空间衍生映射提升秘密恢复率,同时减少负载恢复性,并提出基于机制可解释性的检测方法,提高微调模型中的秘密检测准确率。

详情

展开后加载摘要…

URL PDF HTML 收藏
2508.00923 2026-07-17 cs.LG 版本更新

Addressing Benchmarking Gaps in Large Language Models for Health and Medicine with Dynamic Red-Teaming

超越基准:动态、自动和系统化的红队代理用于可信的医疗语言模型

Jiazhen Pan, Bailiang Jian, Paul Hager, Yundi Zhang, Che Liu, Friederike Jungmann, Hongwei Bran Li, Julian Canisius, Chenyu You, Junde Wu, Jiayuan Zhu, Fenglin Liu, Yuyuan Liu, Niklas Bubeck, Moritz Knolle, Chen, Chen, Christian Wachinger, Zhenyu Gong, Cheng Ouyang, Georgios Kaissis, Benedikt Wiestler, Daniel Rueckert

机构 * Technical University of Munich (TUM)(慕尼黑技术大学) University of Oxford(牛津大学) TUM University Hospital(慕尼黑技术大学医院) Imperial College London(伦敦帝国理工学院) Harvard Medical School(哈佛医学院) Stony Brook University(史泰兹布鲁克大学) Munich Center for Machine Learning (MCML)(慕尼黑机器学习中心) University of Sheffield(谢菲尔德大学)

AI总结 本文提出DAS红队框架,通过动态压力测试揭示医疗语言模型在鲁棒性、隐私、偏见和幻觉方面的潜在风险,发现高静态基准性能与低动态可靠性之间的'基准差距'。

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.18685 2026-07-16 cs.CV cs.RO 版本更新

Beyond Description: Cognitively Benchmarking Fine-Grained Action for Embodied Agents

超越描述:为具身智能体进行细粒度动作的认知基准测试

Dayong Liu, Chao Xu, Weihong Chen, Suyu Zhang, Juncheng Wang, Jiankang Deng, Baigui Sun, Yang Liu

机构 * Zhejiang University(浙江大学) Wolf 1069 b(沃尔夫1069b) Sany Group(三一集团) The Hong Kong Polytechnic University(香港理工大学) Imperial College London(帝国理工学院)

AI总结 本文提出CFG-Bench基准测试,旨在评估具身智能体在物理交互中的细粒度动作能力,揭示现有MLLMs在高层次推理中的不足,并通过监督微调提升其性能。

Comments Accepted to ECCV2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.22415 2026-07-16 cs.CV cs.AI 版本更新

Evidence Recomposition and Predictive Context Residualization for Visual Attribution in Multimodal Large Language Models

多模态大语言模型中用于视觉归因的证据重组与预测上下文残差化

Jiawei Liang, Jianjie Huang, Ruoyu Chen, Xianghao Jiao, Siyuan Liang, Shiming Liu, Xiaochun Cao

机构 * Shenzhen Campus of Sun Yat-sen University(中山大学深圳校区) Zhongguancun Academy(中关村学院) University of Chinese Academy of Sciences(中国科学院大学) Nanyang Technological University(南洋理工大学) Department of Mechanical Engineering, Imperial College London(伦敦帝国理工学院机械工程系)

AI总结 研究多模态大语言模型token级视觉证据难检查问题,提出基于证据重组和预测上下文残差化的ERCR框架,经实验验证该框架能改善目标token视觉证据、减轻上下文干扰,为视觉证据检查提供实用改进。

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.01568 2026-07-15 cs.LG cs.CV cs.IT math.IT q-bio.NC 版本更新

Same Compression Principle, Different Geometry: Rate-Distortion Signatures Dissociate Biological and Artificial Visual Systems

泛化与信息权衡的率-失真签名

Leyla Roksan Caglar, Pedro A. M. Mediano, Baihan Lin

机构 * Windreich Department of AI Human Health, Icahn School of Medicine at Mount Sinai, New York, NY, USA Department of Computing, Imperial College London, London, UK Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, NY, USA Department of Neuroscience, Icahn School of Medicine at Mount Sinai, New York, NY, USA Berkman Klein Center for Internet \& Society, Harvard University, Cambridge, MA, USA

AI总结 本文提出率-失真理论框架,通过斜率和曲率签名分析系统泛化与鲁棒性权衡,揭示生物与人工系统在RD空间中的不同表现。

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.04144 2026-07-15 cs.AI 版本更新

RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories

RippleBench: 利用现有知识库捕捉涟漪效应

Roy Rinberg, Usha Bhalla, Igor Shilov, Flavio P. Calmon, Rohit Gandikota

机构 * Harvard University(哈佛大学) Imperial College London(伦敦帝国学院) Northeastern University(东北大学)

AI总结 提出RippleBench-Maker自动管道,从知识库检索语义邻居生成选择题,评估八种遗忘方法在Llama3-8B-Instruct上的涟漪效应,发现准确率下降随语义距离衰减且跨模型一致。

详情

展开后加载摘要…

URL PDF HTML 收藏
2312.17670 2026-07-15 cs.CV cs.LG q-bio.QM q-bio.TO 版本更新

The TopCoW Challenge -- Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

TopCoW挑战——用于CT和MR血管造影的拓扑感知Willis环分割

Kaiyuan Yang, Fabio Musio, Yihui Ma, Norman Juchler, Johannes C. Paetzold, Rami Al-Maskari, Luciano Höher, Hongwei Bran Li, Ibrahim Ethem Hamamci, Anjany Sekuboyina, Suprosanna Shit, Houjing Huang, Chinmay Prabhakar, Ezequiel de la Rosa, Bastian Wittmann, Diana Waldmannstetter, Florian Kofler, Fernando Navarro, Martin J. Menten, Ivan Ezhov, Daniel Rueckert, Iris N. Vos, Ynte M. Ruigrok, Birgitta K. Velthuis, Hugo J. Kuijf, Pengcheng Shi, Wei Liu, Ting Ma, Maximilian R. Rokuss, Yannick Kirchhoff, Fabian Isensee, Klaus Maier-Hein, Chengcheng Zhu, Huilin Zhao, Philippe Bijlenga, Julien Hämmerli, Catherine Wurster, Laura Westphal, Jeroen Bisschop, Elisa Colombo, Hakim Baazaoui, Hannah-Lea Handelsmann, Andrew Makmur, James Hallinan, Amrish Soundararajan, Benedikt Wiestler, Jan S. Kirschke, Evamaria O. Riedel, Roland Wiest, Emmanuel Montagnon, Laurent Letourneau-Guillon, Kwanseok Oh, Dahye Lee, Orhun Utku Aydin, Adam Hilbert, Jana Rieger, Dimitrios Rallios, Satoru Tanioka, Alexander Koch, Dietmar Frey, Abdul Qayyum, Moona Mazher, Steven Niederer, Nico Disch, Julius C. Holzschuh, Dominic LaBella, Francesco Galati, Daniele Falcetta, Maria A. Zuluaga, Chaolong Lin, Haoran Zhao, Zehan Zhang, Minghui Zhang, Xin You, Hanxiao Zhang, Guang-Zhong Yang, Yun Gu, Sinyoung Ra, Jongyun Hwang, Hyunjin Park, Junqiang Chen, Marek Wodzinski, Henning Müller, Nesrin Mansouri, Florent Autrusseau, Cansu Yalcin, Rachika E. Hamadache, Clara Lisazo, Joaquim Salvi, Adrià Casamitjana, Xavier Lladó, Uma Maria Lal-Trehan Estrada, Valeriia Abramova, Luca Giancardo, Arnau Oliver, Paula Casademunt, Adrian Galdran, Matteo Delucchi, Oscar Camara, Jialu Liu, Haibin Huang, Yue Cui, Zehang Lin, Yusheng Liu, Shunzhi Zhu, Tatsat R. Patel, Adnan H. Siddiqui, Vincent M. Tutino, Maysam Orouskhani, Huayu Wang, Mahmud Mossa-Basha, Yuki Sato, Sven Hirsch, Susanne Wegener, Bjoern Menze

机构 * Department of Quantitative Biomedicine, University of Zurich, Zurich, Switzerland Institute of Computational Life Sciences, Zurich University of Applied Sciences (ZHAW), Waedenswil, Switzerland Department of Neuroradiology, University Hospital of Zurich, Zurich, Switzerland Department of Neurosurgery, Zhongnan Hospital of Wuhan University, Wuhan, China Department of Radiology at Weill Cornell Medicine, Cornell University, New York, USA Institute for Tissue Engineering School of Computation, Information Technology, Technical University of Munich, Germany Athinoula A. Martinos Center for Biomedical Imaging, Harvard Medical School, Boston, USA School of Medicine Health, TUM Klinikum, Technical University of Munich, Germany Munich Center for Machine Learning, Munich, Germany Department of Computing, Imperial College London, London, UK Image Sciences Institute, UMC Utrecht, Utrecht, The Netherlands Department of Neurology Neurosurgery, University Medical Center Utrecht, Utrecht, The Netherlands Department of Radiology, University Medical Center Utrecht, Utrecht, The Netherlands Electronic \& Information Engineering School, Harbin Institute of Technology (Shenzhen), China Peng Cheng Laboratory, Shenzhen, China Division of Medical Image Computing, German Cancer Research Center (DKFZ), Heidelberg, Germany Faculty of Mathematics Computer Science, Heidelberg University, Germany Helmholtz Imaging, German Cancer Research Center, Heidelberg, Germany Data Science School for Health, Karlsruhe/Heidelberg, Germany Learning Group, Department of Radiation Oncology, Heidelberg University Hospital Department of Radiology, University of Washington, Seattle, WA, USA Department of Radiology, Ren Ji Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Department of Clinical Neurosciences, Division of Neurosurgery, Geneva University Hospitals, Geneva, Switzerland Department of Neurology, University Hospital of Zurich, Zurich, Switzerland Department of Physiology, University of Toronto, Canada Department of Neurosurgery, University Hospital of Zurich, Zurich, Switzerland Department of Diagnostic Imaging, National University Hospital, Singapore University of Chicago, USA Department of Diagnostic Interventional Neuroradiology, University Hospital Berne University of Berne, Berne, Switzerland Centre de Recherche du Centre Hospitalier de l’Université de Montréal (CRCHUM), Montréal, Québec, Canada DEEPNOID Inc., Seoul, South Korea Department of Artificial Intelligence, Korea University, Seoul, South Korea Charité Lab for AI in Medicine (CLAIM), Charité Universitätsmedizin Berlin, Berlin, Germany Lung Institute, Faculty of Medicine, Imperial College London, London, UK Centre for Medical Image Computing, Department of Computer Science, University College London, London, UK Department of Radiation Oncology, Duke University Medical Center, Durham, NC, USA Institute of Medical Technology, Peking University Health Science Center, Beijing, China Hangzhou Genlight MedTech Co., Ltd., China Institute of Medical Robotics, Shanghai Jiao Tong University, Shanghai, China Department of Automation, Shanghai Jiao Tong University, Shanghai, China Department of Artificial Intelligence, Sungkyunkwan University, Seoul, South Korea Department of Electrical Computer Engineering, Sungkyunkwan University, Seoul, South Korea Shanghai MediWorks Precision Instruments Co., Ltd., China Institute of Informatics, HES-SO Valais-Wallis, Switzerland Department of Measurement Electronics, AGH University of Krakow, Poland Laboratoire de Thermique et Energie de Nantes (LTeN), Université Nantes, Polytech’Nantes, Nantes, France Research Institute of Computer Vision Center for Precision Health, McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston, USA Physense, BCN-Medtech, Department of Communication Information Technologies, Universitat Pompeu Fabra, Barcelona, Spain Department of Mathematical Modeling Machine Learning, University of Zurich, Zurich, Switzerland Laboratory of Brain Atlas Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences, Beijing, China School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China School of Computer Information Engineering, Xiamen University of Technology, Xiamen, China Vascular Research Center, University at Buffalo, NY, USA Department of Pathology Anatomical Sciences, University at Buffalo, NY, USA Department of Neurosurgery, University at Buffalo, NY, USA LPIXEL Inc., Tokyo, Japan

AI总结 组织TopCoW基准挑战,发布含125对MRA和CTA扫描的注释数据集,参与者提交CoW分割和变体分类算法,经评估,最佳算法在多任务中表现出色,证明CoW分割算法对下游临床应用有可解释性效用。

Comments Summary paper for the TopCoW Challenge: 4 figures, 1 table, and supplementary material in appendix. Accepted for publication in NEJM AI. Datasets and best-performing algorithm Dockers are available at https://zenodo.org/records/15692630 and https://zenodo.org/records/15665435

Journal ref NEJM AI 2026;3(8)

详情

展开后加载摘要…

URL PDF HTML 收藏
2405.19466 2026-07-15 cs.LG stat.ML 版本更新

Active Exploration via Autoregressive Generation of Missing Data

通过自回归生成缺失数据进行主动探索

Tiffany Tianhui Cai, Hongseok Namkoong, Daniel Russo, Kelly W Zhang

机构 * Columbia University(哥伦比亚大学) Imperial College London(帝国理工学院)

AI总结 将在线决策中的不确定性量化和探索问题转化为自回归序列模型的训练与生成,通过预测缺失结果而非潜在参数来建模不确定性,理论证明在线学习可归约为离线下一结果预测,并在新闻推荐中验证有效性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2508.21787 2026-07-14 cs.CL cs.AI 版本更新

PiCSAR: Probabilistic Confidence Selection And Ranking for Reasoning Chains

PiCSAR:基于推理链的概率置信度选择与排序

Joshua Ong Jun Leang, Zheng Zhao, Aryo Pradipta Gema, Sohee Yang, Wai-Chung Kwan, Xuanli He, Wenda Li, Pasquale Minervini, Eleonora Giunchiglia, Shay B. Cohen

机构 * Imperial College London(伦敦帝国学院) University of Edinburgh(爱丁堡大学) UCL(伦敦大学学院)

AI总结 PiCSAR通过联合对数似然度评估推理链和最终答案的置信度,无需训练即可提升大语言模型和推理模型的准确性,在多个基准测试中表现优异。

Journal ref Findings of the Association for Computational Linguistics: ACL 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.02979 2026-07-14 physics.ao-ph cs.LG 版本更新

Parameter estimation for land-surface models using Neural Physics

利用神经物理进行陆面模型参数估计

Ruiyue Huang, Claire E. Heaney, Maarten van Reeuwijk

机构 * Department of Civil and Environmental Engineering, Imperial College London(帝国理工学院土木与环境工程系) Department of Earth Science and Engineering, Imperial College London(帝国理工学院地球科学与工程系) Imperial-X, Imperial College London(帝国理工学院Imperial-X)

AI总结 本文提出一种新型反演方法,通过将数据融入可微物理基础前向模型中,直接优化时间依赖参数,无需推导和维护伴随公式。使用合成数据验证,显示单层土壤温度时间序列无法可靠估计参数,但双层测量可可靠估计参数,同时无法区分潜热和显热通量。

Comments 18 pages, 5 figures, 3 tables

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.13840 2026-07-14 cs.CV 版本更新

MoLingo: Motion-Language Alignment for Text-to-Human Motion Generation

MoLingo:用于文本到运动生成的运动-语言对齐

Yannan He, Garvita Tiwari, Xiaohan Zhang, Pankaj Bora, Tolga Birdal, Jan Eric Lenssen, Gerard Pons-Moll

机构 * University of Tübingen(图宾根大学) Tübingen AI Center(图宾根人工智能中心) Max Planck Institute for Informatics(马克斯·普朗克信息学研究所) Imperial College London(伦敦帝国理工学院) Zuse School ELIZA(Zuse ELIZA 学院)

AI总结 本文提出MoLingo模型,通过在连续潜在空间中去噪生成逼真的人体运动。研究如何构建语义对齐的潜在空间和最佳注入文本条件以提高运动真实性与描述一致性。

Comments Accepted by CVPR 2026. Project page: https://hynann.github.io/molingo/MoLingo.html. Title type fixed, content unchanged

Journal ref Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recogn. (CVPR), 2026, pp. 38387-38398

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.15353 2026-07-14 stat.AP cs.LG stat.ML 版本更新

Reinforcement Learning in the Real World: A Survey of Statistical Challenges and Future Directions

现实世界中的强化学习:统计挑战与未来方向综述

Asim H. Gazi, Yongyi Guo, Daiqi Gao, Ziping Xu, Kelly W. Zhang, Susan A. Murphy

机构 * Department of Computer Science, Harvard University(哈佛大学计算机科学系) Department of Statistics, University of Wisconsin–Madison(威斯康星大学麦迪逊分校统计学系) Department of Statistics, Harvard University(哈佛大学统计学系) School of Data Science and Society, University of North Carolina at Chapel Hill(北卡罗来纳大学教堂山分校数据科学与社会学院) Department of Mathematics, Imperial College London(伦敦帝国理工学院数学系)

AI总结 本文综述现实世界强化学习应用,指出其研究与部署存在差距及两大挑战。将应用框架化为三部分过程,回顾应对统计挑战的进展,涵盖在线、离线方法及持续改进设计,还概述受应用启发的未来研究方向。

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.17502 2026-07-13 cs.AI 版本更新

Towards Shutdownable Agents: Generalizing Stochastic Choice in RL Agents and LLMs

迈向可关闭的智能体:在强化学习智能体和大语言模型中推广随机选择

Carissa Cullen, Harry Garland, Alexander Roman, Louis Thomson, Christos Ziakas, Elliott Thornley

机构 * University of Oxford(牛津大学) University College London(伦敦大学学院) New College of Florida(佛罗里达新学院) Imperial College London(伦敦帝国学院) MIT(麻省理工学院)

AI总结 本文提出DReST奖励函数,通过惩罚重复选择相同长度轨迹来训练智能体在不同轨迹长度间随机选择并有效追求目标,实验表明DReST训练的智能体在未见过的场景中表现出更高的有用性和中立性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.16481 2026-07-10 cs.LG cs.SY eess.SY math.OC 版本更新

Optimal uncertainty bounds for multivariate kernel regression under bounded noise: A Gaussian process-based dual function

有界噪声下多元核回归的最优不确定性界:基于高斯过程的对偶函数

Amon Lahr, Anna Scampicchio, Johannes Köhler, Melanie N. Zeilinger

机构 * Institute for Dynamical Systems and Control, ETH Zurich(动态系统与控制研究所,苏黎世联邦理工学院) Department of Electrical Engineering, Chalmers University of Technology(电气工程系,查尔姆斯理工大学) Department of Mechanical Engineering, Imperial College London(机械工程系,伦敦帝国理工学院)

AI总结 针对有界噪声下再生核希尔伯特空间中的多输出函数,提出一种紧致、确定性的不确定性界,通过无约束对偶公式获得,具有与经典高斯过程置信界相同的结构,便于集成到下游优化中。

Comments Extended version

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.10946 2026-07-10 cs.IT cs.AI cs.LG eess.SP math.IT 版本更新

ToDMA: Large Model-Driven Massive Token Communications for Semantic Multiple Access

ToDMA:用于语义多址接入的大模型驱动海量令牌通信

Li Qiao, Mahdi Boloursaz Mashhadi, Zhen Gao, Robert Schober, Deniz Gündüz

机构 * The University of Hong Kong(香港大学) Beijing Institute of Technology(北京理工大学) University of Surrey(Surrey大学) Friedrich-Alexander-University Erlangen-Nurnberg(埃森哲-亚琛工业大学) Imperial College London(伦敦帝国理工学院)

AI总结 本文提出ToDMA,一种大模型驱动的语义多址接入方案,将无源随机接入与上下文感知令牌处理结合,用于海量令牌通信。通过压缩感知检测令牌及估计信道状态信息,利用上下文模型恢复冲突令牌,能降低接入延迟并保持令牌恢复与语义重建质量。

Comments Submitted to IEEE journals

详情

展开后加载摘要…

URL PDF HTML 收藏
2603.01730 2026-07-09 cs.LG 版本更新

Decentralized Federated Learning by Partial Message Exchange

通过部分消息交换进行去中心化联邦学习

Shan Sha, Shenglong Zhou, Xin Wang, Lingchen Kong, Geoffrey Ye Li

机构 * School of Mathematics and Statistics, Beijing Jiaotong University(北京交通大学数学与统计学学院) Department of Electrical and Electronic Engineering, Faculty of Engineering, Imperial College London(帝国理工学院伦敦分校电子与电气工程系)

AI总结 研究去中心化联邦学习面临的挑战,提出PaME算法,通过允许相邻节点仅交换随机选择的稀疏坐标,降低通信成本、保护隐私且不牺牲准确性,经分析在温和假设下线性收敛,数值实验证明性能优越。

详情

展开后加载摘要…

URL PDF HTML 收藏
2411.01683 2026-07-09 cs.CV 版本更新

ROAD-Waymo: A Large-Scale Action Awareness Dataset for Autonomous Driving

ROAD-Waymo:一个用于自动驾驶的大规模动作感知数据集

Salman Khan, Izzeddin Teeti, Reza Javanmard Alitappeh, Mihaela C. Stoian, Eleonora Giunchiglia, Gurkirt Singh, Andrew Bradley, Fabio Cuzzolin

机构 * Oxford Brookes University(奥克斯福德布鲁克斯大学) MAZUST University of Oxford(牛津大学) Imperial College London(伦敦帝国学院) ETH Zurich(苏黎世联邦理工学院)

AI总结 本文提出ROAD-Waymo数据集,用于道路场景中代理、动作等检测技术开发与基准测试,比现有数据集更大更具挑战性,含大量标注数据,通过新注释管道增强完整性,还能解决不同国家道路场景的域适应问题。

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.18419 2026-07-08 cs.CV cs.AI 版本更新

Geometry-Aware Uncertainty Coresets for Robust Visual In-Context Learning in Histopathology

面向几何的不确定性聚类用于病理学中鲁棒的视觉上下文学习

Franciskus Xaverius Erick, Johanna Paula Müller, Bernhard Kainz

机构 * FAU Erlangen-Nürnberg, Erlangen, DE(埃尔兰根-纽伦堡大学) Department of Computing, Imperial College London, London, UK(伦敦帝国理工学院计算机系)

AI总结 本文提出GAUC,一种无需训练的聚类选择方法,直接在预训练的多模态嵌入空间中操作,通过优化三个目标提升视觉上下文学习的鲁棒性、准确性和校准性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.12100 2026-07-08 cs.CV 版本更新

Did Models Learn Sufficiently? Attribution-Guided Training via Subset-Selected Counterfactual Augmentation

模型是否充分学习?通过子集选择的反事实增强进行归因引导训练

Yannan Chen, Ruoyu Chen, Wei Wang, Bin Zeng, Jinke Li, Shiming Liu, Qunli Zhang, Yaowei Wang, Xiaochun Cao

机构 * School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University(中山大学深圳校区信息科学与技术学院) PCL Institute of Information Engineering, CAS(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) Tianjin University(天津大学) Lanzhou University(兰州大学) Imperial College London(伦敦帝国理工学院)

AI总结 针对模型仅依赖有限充分原因导致泛化性差的问题,提出子集选择反事实增强(SS-CA),通过基于LIMA归因的反事实生成和数据增强,改善模型因果学习,提升分布内和分布外性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.13305 2026-07-07 cs.LG math.DS physics.chem-ph 版本更新

MPINeuralODE: Multiple-Initial-Condition Physics-Informed Neural ODEs for Globally Consistent Dynamical System Learning

MPINeuralODE:多初始条件物理信息神经ODE用于全局一致的动力系统学习

Lake Yang, Antonio Malpica-Morales, Frank Ioannis Papadakis Wood, Serafim Kalliadasis

机构 * Department of Chemical Engineering, Imperial College London(帝国理工学院伦敦校区化学工程系)

AI总结 本文提出MPINeuralODE,结合软物理信息残差和多初始条件多步学习策略,有效提升动力系统学习的泛化能力与稳定性,实验显示其在样本外误差、长时域稳定性及哈密顿漂移方面表现优异。

详情

展开后加载摘要…

URL PDF HTML 收藏
2511.00602 2026-07-07 cs.CL 版本更新

OpenSIR: Open-Ended Self-Improving Reasoner

OpenSIR: 开放式自我改进推理器

Wai-Chung Kwan, Joshua Ong Jun Leang, Pavlos Vougiouklis, Jeff Z. Pan, Marco Valentino, Pasquale Minervini

机构 * University of Edinburgh(爱丁堡大学) Imperial College London(伦敦帝国理工学院)

AI总结 提出OpenSIR框架,通过多样性奖励和难度校准实现开放式自我对弈,无需外部验证器或标注数据,在七个数学基准上平均提升指令模型3.6分、推理模型3.1分,且唯一能泛化到通用推理。

详情

展开后加载摘要…

URL PDF HTML 收藏
2606.04265 2026-07-03 math.OC cs.LG cs.NA math.NA 版本更新

Nonlocal Mean Field Schrödinger Bridge with Learned Interactions

具有学习相互作用的非局部平均场薛定谔桥

Daisuke Inoue, Dante Kalise, Mathieu Laurière

机构 * Department of Mathematics, Imperial College London(伦敦帝国学院数学系) Shanghai Frontiers Science Center of Artificial Intelligence and Deep Learning(上海前沿人工智能与深度学习科学中心) NYU-ECNU Institute of Mathematical Sciences, NYU Shanghai(纽约大学上海数学科学研究所)

AI总结 本文提出一种使用神经网络代理近似非局部相互作用的平均场薛定谔桥方法,将推理时的每步计算成本从二次降低到线性,并推导了代理误差传播的稳定性界限。

Comments 32 pages, 15 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.02546 2026-07-03 cs.CV cs.LG 版本更新

RGB-Pointmap Pretraining for Unified 3D Scene Understanding

对比语言-彩色点图预训练用于统一3D场景理解

Ye Mao, Weixun Luo, Ranran Huang, Junpeng Jing, Krystian Mikolajczyk

机构 * Imperial College London(帝国理工学院伦敦分校)

AI总结 提出UniScene3D,一种基于Transformer的编码器,通过多视图彩色点图联合建模外观和几何,并引入跨视图几何对齐和接地视图对齐,在少样本和任务特定微调中达到SOTA。

Comments 19 Pages, ECCV 2026 Accepted

详情

展开后加载摘要…

URL PDF HTML 收藏
2411.16956 2026-07-03 eess.IV cs.AI cs.CV 版本更新

Contrastive Deep Learning Reveals Age Biomarkers in Histopathological Skin Biopsies

对比深度学习揭示组织病理学皮肤活检中的年龄生物标志物

Kaustubh Chakradeo, Pernille Nielsen, Lise Mette Rahbek Gjerdrum, Gry Sahl Hansen, David A Duchêne, Laust H Mortensen, Majken K Jensen, Samir Bhatt

机构 * University of Copenhagen, Section of Epidemiology, Department of Public Health(哥本哈根大学流行病学系,公共卫生系) Technical University of Denmark, Department of Applied Mathematics and Computer Science(丹麦技术大学应用数学与计算机科学系) Department of Pathology, Copenhagen University Hospital- Zealand University Hospital(哥本哈根大学医院- Zealand大学医院病理科) Department of Clinical Medicine, University of Copenhagen(哥本哈根大学临床医学系) Danmarks Statistik(丹麦统计局) Imperial College London(伦敦帝国理工学院)

AI总结 使用对比深度学习,仅凭皮肤活检图像即可确定个体年龄,并构建新的衰老生物标志物,预测死亡率和慢性年龄相关疾病患病率。

Comments 20 pages, 5 tables, 5 figures Under review: npj Digital Medicine

详情

展开后加载摘要…

URL PDF HTML 收藏
2606.29963 2026-07-02 cs.CV cs.CR 版本更新

Explainability-Aware Frustum Attack: Exposing Structural Vulnerabilities in LiDAR-Based 3D Object Detectors

可解释性感知的视锥攻击:揭示基于LiDAR的3D目标检测器中的结构脆弱性

Chengzeng You, Binbin Xu, Soteris Demetriou

机构 * Imperial College London(帝国理工学院伦敦分校)

AI总结 提出可解释性引导的对抗分析方法,通过SALL方法生成通用显著性图,并设计EFA攻击,仅扰动最关键的视锥区域,在KITTI和nuScenes上使检测召回率下降超15个百分点,同时减少25-50%的扰动视锥数。

Comments European Conference on Computer Vision (ECCV), September 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2604.06817 2026-07-02 cs.CL 版本更新

SemEval-2026 Task 9: Detecting Multilingual, Multicultural and Multievent Online Polarization

SemEval-2026任务9:多语言、多文化、多事件在线极化检测

Usman Naseem, Robert Geislinger, Juan Ren, Sarah Kohail, Rudy Garrido Veliz, P Sam Sahil, Yiran Zhang, Marco Antonio Stranisci, Idris Abdulmumin, Özge Alaçam, Cengiz Acartürk, Aisha Jabr, Saba Anwar, Abinew Ali Ayele, Elena Tutubalina, Aung Kyaw Htet, Xintong Wang, Surendrabikram Thapa, Tanmoy Chakraborty, Dheeraj Kodati, Sahar Moradizeyveh, Firoj Alam, Ye Kyaw Thu, Shantipriya Parida, Ihsan Ayyub Qazi, Lilian Wanzare, Nelson Odhiambo Onyango, Clemencia Siro, Ibrahim Said Ahmad, Adem Chanie Ali, Martin Semmann, Chris Biemann, Shamsuddeen Hassan Muhammad, Seid Muhie Yimam

机构 * Macquarie University(麦考瑞大学) University of Hamburg(汉堡大学) Zayed University(扎耶德大学) HKBK College of Engineering(HKBK工程学院) University of Turin(都灵大学) aequa-tech University of Pretoria(比勒陀利亚大学) Bielefeld University(比勒费尔德大学) Jagiellonian University(雅盖隆大学) Bahir Dar University(巴赫达尔大学) AIRI KFU(喀山联邦大学) HSE University(高等经济大学) Virginia Tech(弗吉尼亚理工大学) IIT Delhi(印度理工学院德里分校) ABV-IIITM(ABV-印度信息技术与管理学院) Qatar Computing Research Institute(卡塔尔计算研究所) Hamad Bin Khalifa University(哈马德·本·哈利法大学) Language Understanding Lab., Myanmar(缅甸语言理解实验室) AMD Silo AI Lahore University of Management Sciences(拉合尔管理科学大学) Maseno University(马塞诺大学) Centrum Wiskunde & Informatica(数学与计算机科学中心) Bayero University Kano(卡诺巴耶罗大学) Northeastern University(东北大学) Imperial College London(伦敦帝国学院)

AI总结 本文介绍了SemEval-2026任务9,旨在检测在线极化,涵盖22种语言,包含超过11万标注实例。任务包含三个子任务,吸引全球1000多参与者和10000多提交,最终有67支队伍提交了系统描述论文。

详情

展开后加载摘要…

URL PDF HTML 收藏
2503.13445 2026-07-02 cs.CL cs.AI 版本更新

Verbosity Tradeoffs and the Impact of Scale on the Faithfulness of LLM Self-Explanations

冗长性权衡与规模对LLM自我解释忠实度的影响

Noah Y. Siegel, Nicolas Heess, Maria Perez-Ortiz, Oana-Maria Camburu

机构 * Google DeepMind(谷歌DeepMind) Centre for AI, University College London(伦敦大学学院人工智能中心) Imperial College London(伦敦帝国学院) University College London(伦敦大学学院)

AI总结 本文分析13个模型家族共75个模型的反事实忠实度,提出phi-CCT和F-AUROC两个新指标,发现更大更强的模型在所有指标上更忠实。

Comments ICLR 2026 Workshop on Principled Design for Trustworthy AI - Interpretability, Robustness, and Safety across Modalities 67 pages, 13 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.16399 2026-07-01 cs.CV cs.LG 版本更新

Stable and Near-Reversible Diffusion ODE Solvers for Image Editing

稳定且近可逆的图像编辑扩散ODE求解器

Barbora Barancikova, Daniil Shmelev, Cristopher Salvi

机构 * Department of Computing, Imperial College London, London, United Kingdom(帝国理工学院伦敦分校计算机系,伦敦,英国) Department of Mathematics, Imperial College London, London, United Kingdom(帝国理工学院伦敦分校数学系,伦敦,英国)

AI总结 本文提出近可逆Runge-Kutta方法以提升图像编辑的稳定性与精度,平衡可逆性与数值稳定性,保留背景保真优势。

Comments ICML 2026 Workshop on Structured Probabilistic Inference & Generative Modeling (SPIGM)

详情

展开后加载摘要…

URL PDF HTML 收藏
2508.09156 2026-07-01 cs.LG cs.AI stat.AP 版本更新

Physics-Constrained Fine-Tuning of Flow-Matching Models for Generation and Inverse Problems

面向生成和逆问题的流匹配模型的物理约束微调

Jan Tauberschmidt, Sophie Fellenz, Sebastian J. Vollmer, Andrew B. Duncan

机构 * German Research Center for Artificial Intelligence (DFKI)(德国人工智能研究中心) Department of Computer Science, University of Kaiserslautern–Landau (RPTU)(凯泽斯劳滕-兰道大学计算机科学系) Department of Mathematics, Imperial College London(伦敦帝国理工学院数学系)

AI总结 提出一种微调流匹配生成模型以强制执行物理约束并解决科学系统逆问题的框架,通过微分后训练最小化偏微分方程弱形式残差,并联合优化可学习潜参数预测器,实现物理一致性与潜系数准确恢复。

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.17246 2026-07-01 cs.CV 版本更新

SPFSplatV2: Efficient Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Views

SPFSplatV2: 基于稀疏视图的高效自监督无姿态3D高斯泼溅

Ranran Huang, Krystian Mikolajczyk

机构 * Department of Electrical and Electronic Engineering, Imperial College London(伦敦帝国理工学院电气与电子工程系)

AI总结 提出SPFSplatV2,一种从稀疏多视图图像进行3D高斯泼溅的前馈框架,无需真实姿态,通过共享特征提取、掩码注意力机制和重投影损失实现高效姿态估计和新视图合成,在域内外均达到最先进性能。

详情

展开后加载摘要…

URL PDF HTML 收藏