arXivDaily arXiv每日学术速递 周一至周五更新

高校专区

Northeastern University(东北大学)

2026-01-13 至 2026-01-13 共收录 12
2601.07645 2026-01-13 cs.CL

PlaM: Training-Free Plateau-Guided Model Merging for Better Visual Grounding in MLLMs

PlaM: 无需训练的高原引导模型融合以提升多模态大语言模型的视觉语义

Zijing Wang, Yongkang Liu, Mingyang Wang, Ercong Nie, Deyuan Chen, Zhengjie Zhao, Shi Feng, Daling Wang, Xiaocui Yang, Yifei Zhang, Hinrich Schütze

机构 * Northeastern University, China(东北大学) CIS, LMU Munich, Germany(慕尼黑大学计算机学院) Munich Center for Machine Learning (MCML), Germany(慕尼黑机器学习中心)

AI总结 PlaM通过无需训练的高原引导模型融合方法,提升多模态大语言模型的视觉语义表现。

Comments under review

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.07507 2026-01-13 cs.CL

High-Rank Structured Modulation for Parameter-Efficient Fine-Tuning

高秩结构调制用于参数高效微调

Yongkang Liu, Xing Li, Mengjie Zhao, Shanru Zhang, Zijing Wang, Qian Li, Shi Feng, Feiliang Ren, Daling Wang, Hinrich Schütze

机构 * Northeastern University, China(东北大学) CIS, LMU Munich, Germany(慕尼黑大学计算机科学系) Shandong University, China(山东大学) Munich Center for Machine Learning (MCML), Germany(慕尼黑机器学习中心)

AI总结 SMoA通过高秩结构调制在减少可训练参数的同时提升模型代表能力,优于LoRA在多个任务上表现更优。

Comments under review

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.07423 2026-01-13 cs.CL

SAD: A Large-Scale Strategic Argumentative Dialogue Dataset

SAD:一个大规模的战略辩论对话数据集

Yongkang Liu, Jiayang Yu, Mingyang Wang, Yiqun Zhang, Ercong Nie, Shi Feng, Daling Wang, Kaisong Song, Hinrich Schütze

机构 * Northeastern University, China(东北大学) CIS, LMU Munich, Germany(慕尼黑莱茵-穆尔大学认知科学研究所) Munich Center for Machine Learning (MCML), Germany(慕尼黑机器学习中心) Alibaba Group, Hangzhou, China(阿里巴巴集团)

AI总结 SAD数据集旨在通过大规模战略辩论对话数据支持更深入的论证对话建模,包含392,822个示例,标注五种策略类型,并测试多种预训练模型。

Comments under review

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.00245 2026-01-13 cs.NE cs.IT cs.LG math.IT

Modern Neuromorphic AI: From Intra-Token to Inter-Token Processing

现代神经形态AI:从内词到跨词处理

Osvaldo Simeone

机构 * Intelligent Networked Systems Institute (INSI), Northeastern University London(智能网络系统研究所(INSI),伦敦诺思安普顿大学)

AI总结 本文探讨了神经形态AI中内词与跨词处理的区别,分析了神经形态模型、状态空间模型和Transformer架构的联系,并回顾了训练方法。

详情

展开后加载摘要…

URL PDF HTML 收藏
2512.07873 2026-01-13 cs.LG cs.AI

Advancing time series completion via RFAMoE and MDFF

通过RFAMoE和MDFF推进时间序列补全

Ci Zhang, Huayu Li, Changdi Yang, Jiangnan Xia, Yanzhi Wang, Xiaolong Ma, Jin Lu, Ao Li, Geng Yuan

机构 * University of Georgia(佐治亚大学) University of Arizona(亚利桑那大学) Northeastern University(东北大学)

AI总结 本文提出基于MoE的噪声估计器和RFAMoE、MDFF模块,通过自适应接收域和并行信号融合提升医疗时间序列补全性能。

详情

展开后加载摘要…

URL PDF HTML 收藏
2501.13772 2026-01-13 cs.SD cs.AI cs.LG cs.MM eess.AS

Jailbreak-AudioBench: In-Depth Evaluation and Analysis of Jailbreak Threats for Large Audio Language Models

Jailbreak-AudioBench: 对大型音频语言模型中 jailbreak 威胁的深入评估与分析

Hao Cheng, Erjia Xiao, Jing Shao, Yichi Wang, Le Yang, Chao Shen, Philip Torr, Jindong Gu, Renjing Xu

机构 * Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)) University of Oxford(牛津大学) Xi’an Jiaotong University(西安交通大学) Hong Kong University of Science and Technology(香港科技大学) Northeastern University(东北大学) Beijing University of Technology(北京理工大学)

AI总结 Jailbreak-AudioBench 通过构建工具箱、数据集和基准,深入评估大型音频语言模型中 jailbreak 威胁,并促进安全防护机制的发展。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06799 2026-01-13 cs.CL cs.AI

CIRAG: Construction-Integration Retrieval and Adaptive Generation for Multi-hop Question Answering

CIRAG:多跳问答中的构造-整合检索与自适应生成

Zili Wei, Xiaocui Yang, Yilin Wang, Zihan Wang, Weidong Bao, Shi Feng, Daling Wang, Yifei Zhang

机构 * Northeastern University(东北大学)

AI总结 CIRAG通过构造-整合模块和自适应生成模块,提升多跳问答的准确性和鲁棒性。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06180 2026-01-13 cs.LG cs.AI cs.CL

MixDPO: Modeling Preference Strength for Pluralistic Alignment

MixDPO:建模偏好强度以实现多元对齐

Saki Imai, Pedram Heydari, Anthony Sicilia, Asteria Kaeberlein, Katherine Atwell, Malihe Alikhani

机构 * Northeastern University(东北大学) Johns Hopkins University(约翰霍普金斯大学) West Virginia University(西弗吉尼亚大学)

AI总结 MixDPO通过建模偏好强度差异,提升多样的偏好对齐性能,同时保持子群体偏好,适用于高异质性场景。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06103 2026-01-13 cs.LG cs.AI

The Impact of Post-training on Data Contamination

训练后阶段对数据污染的影响

Muhammed Yusuf Kocyigit, Caglar Yildirim

机构 * Boston University(波士顿大学) Northeastern University(东北大学)

AI总结 研究发现数据污染在训练后阶段会引发性能波动,但通过SFT和GRPO方法可缓解,且模型规模越大,污染影响越显著。

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.06064 2026-01-13 cs.CY cs.AI cs.MA

Socio-technical aspects of Agentic AI

群体技术视角下的代理AI

Praveen Kumar Donta, Alaa Saleh, Ying Li, Shubham Vaishnav, Kai Fang, Hailin Feng, Yuchao Xia, Thippa Reddy Gadekallu, Qiyang Zhang, Xiaodan Shi, Ali Beikmohammadi, Sindri Magnússon, Ilir Murturi, Chinmaya Kumar Dehury, Marcin Paprzycki, Lauri Loven, Sasu Tarkoma, Schahram Dustdar

机构 * Department of Computer and Systems Sciences, Stockholm University(斯德哥尔摩大学计算机与系统科学系) Center for Ubiquitous Computing, University of Oulu(奥卢大学无处不在计算中心) College of Computer Science and Engineering, Northeastern University(东北大学计算机科学与工程学院) Zhejiang A\&F University, Hangzhou(浙江工业大学之江学院) School of Computer Science, Peking University(北京大学计算机科学学院) Department of Mechatronics, University of Prishtina(普里什蒂纳大学机电系) Department of Computer Science, IISER Berhampur(伯尔哈普尔IISER计算机科学系) Systems Research Institute Polish Academy of Sciences(波兰科学院系统研究所) Department of Computer Science, University of Helsinki(赫尔辛基大学计算机科学系)

AI总结 本文从社会技术视角探讨代理AI,分析其技术组件与社会背景的关联,揭示伦理挑战及未来研究方向。

Comments Dear Reviewer, please note that this is not survey/review or position paper. This paper introduced new framework (MAD-BAD-SAD Framework) for Socio-technical aspects of Agentic AI, Ethical considerations, which is very important to consider beside technical development

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.09482 2026-01-13 cs.DB cs.LG

Database Views as Explanations for Relational Deep Learning

数据库视图作为关系深度学习的解释

Agapi Rissaki, Ilias Fountalis, Wolfgang Gatterbauer, Benny Kimelfeld

机构 * Northeastern University(东北大学)

AI总结 本文提出了一种基于视图定义的关系深度学习解释框架,通过可学习掩码实现模型特定的解释方法,提升了解释效率和质量。

详情

展开后加载摘要…

URL PDF HTML 收藏
2508.13021 2026-01-13 cs.AI cs.CL

Empirical Analysis of Decoding Biases in Masked Diffusion Models

掩码扩散模型中解码偏见的实证分析

Pengcheng Huang, Tianming Liu, Zhenghao Liu, Yukun Yan, Shuo Wang, Tong Xiao, Zulong Chen, Maosong Sun

机构 * School of Computer Science and Engineering, Northeastern University, China(东北大学计算机科学与工程学院) Department of Computer Science and Technology, Institute for AI, Tsinghua University, China(清华大学人工智能研究院计算机科学与技术系) Alibaba Group, Hangzhou, China(阿里巴巴集团)

AI总结 本文通过实证分析揭示了掩码扩散模型中注意力漂浮现象,揭示其浅层结构感知与深层内容聚焦的注意力机制,证明其在知识密集型任务中性能优于自回归模型。

Comments 22 pages,17 figures

详情

展开后加载摘要…

URL PDF HTML 收藏