检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Chen, Xinyue Ren, Yazhou Xu, Jie Lin, Fangfei Pu, Xiaorong Yang, Yang School of Computer Science and Engineering University of Electronic Science and Technology of China China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China China

Recently, federated multi-view clustering (FedMVC) has emerged to explore cluster structures in multi-view data distributed on multiple clients. Existing approaches often assume that clients are isomorphic and all of them belong to either single-view clients or multi-view clients. Despite their success, these methods also present limitations when dealing with practical FedMVC scenarios involving heterogeneous hybrid views, where a mixture of both single-view and multi-view clients exhibit varying degrees of heterogeneity. In this paper, we propose a novel FedMVC framework, which concurrently addresses two challenges associated with heterogeneous hybrid views, i.e., client gap and view gap. To address the client gap, we design a local-synergistic contrastive learning approach that helps single-view clients and multi-view clients achieve consistency for mitigating heterogeneity among all clients. To address the view gap, we develop a global-specific weighting aggregation method, which encourages global models to learn complementary features from hybrid views. The interplay between local-synergistic contrastive learning and global-specific weighting aggregation mutually enhances the exploration of the data cluster structures distributed on multiple clients. Theoretical analysis and extensive experiments demonstrate that our method can handle the heterogeneous hybrid views in FedMVC and outperforms state-of-the-art methods. The code is available at https://***/5Martina5/FMCSC. © 2024, CC BY-NC-SA.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Human-in-the-loop Robotic Grasping using BERT Scene Representation 29

Human-in-the-loop Robotic Grasping using BERT Scene Represen...

引用

29th International Conference on Computational Linguistics, COLING 2022

作者： Song, Yaoxian Sun, Penglei Fang, Pengfei Yang, Linyi Xiao, Yanghua Zhang, Yue School of Computer Science Fudan University China School of Engineering Westlake University China School of Engineering Australian National University Australia Institute of Advanced Technology Westlake Institute for Advanced Study China

Current NLP techniques have been greatly applied in different domains. In this paper, we propose a human-in-the-loop framework for robotic grasping in cluttered scenes, investigating a language interface to the grasping process, which allows the user to intervene by natural language commands. This framework is constructed on a state-of-the-art grasping baseline, where we substitute a scene-graph representation with a text representation of the scene using BERT. Experiments on both simulation and physical robot show that the proposed method outperforms conventional object-agnostic and scene-graph based methods in the literature. In addition, we find that with human intervention, performance can be significantly improved. Our dataset and code are available on our project website © 2022 Proceedings - International Conference on Computational Linguistics, COLING. All rights reserved.

关键词： Graphic methods

来源：评论

学校读者我要写书评

暂无评论

Information entropy peaks clustering using dynamic reverse nearest neighbor sequence and 3D decision graph

引用

Expert Systems with Applications 2025年 288卷

作者： Lu, Jianyun Chen, Zhong Shao, Junming Wu, Chunling School of Computer Science and Engineering University of Electronic Science and Technology of China Sichuan Chengdu China School of Artificial Intelligence and Big Data Chongqing Polytechnic University of Electronic Technology Chongqing China School of Computer Science Peking University Beijing China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China China University of Electronic Science and Technology of China China

Density Peaks Clustering has been widely studied due to its ability to detect clusters with arbitrary shapes and sizes. However, inappropriate density measures can result in the detection of multiple density peaks within a single cluster, leading to a degeneration of the availability of density peaks clustering. In this paper, we propose a new concept, information entropy peaks, which is robust to variable densities and noise and can accurately identify cluster representatives. To achieve this, we introduce a dynamic reverse nearest neighbor sequence to compute information entropy for each data point. We then develop a 3D decision graph that utilizes information entropy, Euclidean distance, and cosine distance to identify potential cluster representatives. The cosine distance is calculated based on the dynamic reverse nearest neighbor sequence and can measure distance in the direction of space. Finally, a nearest neighbor propagation algorithm is applied to form clusters, initialized with cluster representatives. Our method is effective in identifying cluster representatives and is robust to variable densities and noise. Comprehensive experiments on both synthetic and real-world datasets have demonstrated that our method produces better clustering results compared to many state-of-the-art clustering algorithms. © 2025 Elsevier Ltd

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

Magnon corner states in twisted bilayer honeycomb magnets

引用

Physical Review B 2023年第2期107卷 L020404-L020404页

作者： Chun-Bo Hua Feiping Xiao Zheng-Rong Liu Jin-Hua Sun Jin-Hua Gao Chui-Zhen Chen Qingjun Tong Bin Zhou Dong-Hui Xu School of Electronic and Information Engineering Hubei University of Science and Technology Xianning 437100 China Laboratory of Optoelectronic Information and Intelligent Control Hubei University of Science and Technology Xianning 437100 China School of Physics and Electronics Hunan University Changsha 410082 China Department of Physics Hubei University Wuhan 430062 China Department of Physics Ningbo University Ningbo 315211 China School of Physics and Wuhan National High Magnetic Field Center Huazhong University of Science and Technology Wuhan 430074 China Institute for Advanced Study and School of Physical Science and Technology Soochow University Suzhou 215006 China Department of Physics and Chongqing Key Laboratory for Strongly Coupled Physics Chongqing University Chongqing 400044 People's Republic of China Center of Quantum Materials and Devices Chongqing University Chongqing 400044 People's Republic of China

The study of symmetry-protected topological phases of matter has been extended from fermionic electron systems to various bosonic systems. Bosonic topological magnon phases in magnetic materials have received much attention because of their exotic uncharged topologically protected boundary modes and the potential for dissipationless magnonics and spintronic applications. Here, we establish twisted bilayer honeycomb magnets as a platform for hosting second-order topological magnon insulators (SOTMIs) without fine-tuning. We employ a simple, minimal Heisenberg spin model to describe misaligned bilayer sheets of honeycomb ferromagnetic magnets with a large commensurate twist angle. We found that the higher-order topology in this bilayer system shows a significant dependence on the interlayer exchange coupling. The SOTMI, featuring topologically protected magnon corner states that go beyond the conventional bulk-boundary correspondence, appears for ferromagnetic interlayer couplings, while the twisted bilayer exhibits a nodal phase in the case of antiferromagnetic interlayer coupling. At last, relevance to twisted bilayer CrI3 is also discussed.

关键词： Magnons Spin waves Topological phases of matter Topological materials

来源：评论

学校读者我要写书评

暂无评论

Bridging the User-side Knowledge Gap in Knowledge-aware Recommendations with Large Language Models 39

Bridging the User-side Knowledge Gap in Knowledge-aware Reco...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Hu, Zheng Li, Zhe Jiao, Ziyun Nakagawa, Satoshi Deng, Jiawen Cai, Shimin Zhou, Tao Ren, Fuji School of Computer Science and Engineering University of Electronic Science and Technology of China China Graduate School of Information Science and Technology University of Tokyo Japan Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China China

ISBN: (纸本)157735897X

In recent years, knowledge graphs have been integrated into recommender systems as item-side auxiliary information, enhancing recommendation accuracy. However, constructing and integrating structural user-side knowledge remains a significant challenge due to the improper granularity and inherent scarcity of user-side features. Recent advancements in Large Language Models (LLMs) offer the potential to bridge this gap by leveraging their human behavior understanding and extensive real-world knowledge. Nevertheless, integrating LLM-generated information into recommender systems presents challenges, including the risk of noisy information and the need for additional knowledge transfer. In this paper, we propose an LLM-based user-side knowledge inference method alongside a carefully designed recommendation framework to address these challenges. Our approach employs LLMs to infer user interests based on historical behaviors, integrating this user-side information with item-side and collaborative data to construct a hybrid structure: the Collaborative Interest Knowledge Graph (CIKG). Furthermore, we propose a CIKG-based recommendation framework that includes a user interest reconstruction module and a cross-domain contrastive learning module to mitigate potential noise and facilitate knowledge transfer. We conduct extensive experiments on three real-world datasets to validate the effectiveness of our method. Our approach achieves state-of-the-art performance compared to competitive baselines, particularly for users with sparse interactions. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing IoMT Security Using Large Multimodal Models

引用

IEEE Network 2025年

作者： Jiao, Yan Razaq, Mian Muaz Peng, Limei Ho, Pin-Han University of Electronic Science and Technology of China Shenzhen Institute for Advanced Study Guangdong China Kyungpook National University School of Computer Science and Engineering Daegu41566 Korea Republic of

This paper explores the potential of Large Multi-modal Models (LMMs) in enhancing the security of Internet of Medical Things (IoMT) networks. As IoMT systems become more complex and generate vast amounts of sensitive data, traditional security measures are increasingly insufficient to address evolving cyber threats. The ability of LMMs to process and integrate diverse data types, such as sensor data, audio, and video, offers a powerful approach to detecting and mitigating security risks in real time. By focusing on key applications like anomaly detection through traffic prediction and improving Physical Layer Security (PLS), this work demonstrates how LMMs can significantly enhance the resilience of IoMT networks against a wide range of cyber threats. The integration of multimodal data enables more accurate traffic predictions and allows adaptive responses to emerging security challenges. Additionally, LMMs contribute to strengthening PLS by identifying vulnerabilities in communication channels and optimizing resource allocation. Through dynamic data fusion techniques, LMMs ensure that IoMT systems can respond to threats in real-time, while maintaining high security and minimizing impact on system performance. © 1986-2012 IEEE.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

SimiDTR: Deep Trajectory Recovery with Enhanced Trajectory Similarity 28th

SimiDTR: Deep Trajectory Recovery with Enhanced Trajectory ...

引用

28th International Conference on Database Systems for Advanced Applications, DASFAA 2023

作者： Zhang, Yupu Deng, Liwei Zhao, Yan Chen, Jin Xie, Jiandong Zheng, Kai University of Electronic Science and Technology of China Chengdu China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China Chengdu China Department of Computer Science Aalborg University Aalborg Denmark Cloud Database Innovation Lab of Cloud BU Huawei Technologies Co. Ltd. Chengdu China

ISBN: (纸本)9783031306365

The pervasiveness of GPS-equipped smart devices and the accompanying deployment of sensing technologies generates increasingly massive amounts of trajectory data that cover applications such as personalized location recommendation and urban transportation planning. Trajectory recovery is of utmost importance for incomplete trajectories (resulted from the constraints of devices and environment) to enable their completeness and reliability. To achieve effective trajectory recovery, we propose a novel trajectory recovery framework, namely Deep Trajectory Recovery with enhanced trajectory Similarity (SimiDTR), which is capable of contending with the complex mobility regularity found in trajectories in continuous space. In particular, we design a rule-based information extractor to extract the spatial information related to an incomplete trajectory, which is then fed into a deep model based on attention mechanism to generate a tailored similar trajectory for the incomplete trajectory. Finally, we use a deep neural network model to recover the incomplete trajectory with the blessing of its similar trajectory. An extensive empirical study with real data offers evidence that the framework is able to advance the state of the art in terms of effectiveness for trajectory recovery, especially in scenes with sparse trajectory data. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Recovery

来源：评论

学校读者我要写书评

暂无评论

INAS: Incremental Neural Architecture Search 27

INAS: Incremental Neural Architecture Search

引用

27th International Conference on Automation and Computing, ICAC 2022

作者： Zhou, Qin Cao, Jiajia Yin, Ling Zhang, Fei Li, Yun Dongguan University of Technology Computer Science and Technology DongGuan China Dongguan University of Technology School of Mechanical Engineering DongGuan China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China Shenzhen China

ISBN: (数字)9781665498074

ISBN: (纸本)9781665498074

Sharing the weights of Super-Net is a common method to improve the efficiency of model performance evaluation in Neural architecture search (NAS). Existing weight-sharing NAS typically uses a fixed Super-Net to be search space, but the Super-Net design relies on the valuable experience of experts for different tasks. When the training dataset or metric is changed, the expert has to design again to obtain optimized architecture. Therefore, we argue that dynamic Super-Net is beneficial to discover promising architecture with less manual intervention during the search stage. Furthermore, we observed that most evolved NAS algorithms ignore the potential connection between the input node and neural operator. This paper describes an incremental neural architecture search (INAS) method that searches for an outstanding neural network architecture by dynamically increasing Super-Net size. And we proposed an improved Heredity Algorithm (HA) as a search strategy, in which overwriting operator is used to process input node and neural operator coding. The overwriting operator can further preserve the outstanding individual genes to offspring and accelerate search convergence. Finally, we performed the experiment on CIFARIO. The search model from our INAS method achieves the highest accuracy 97.97%, beyond the state-of-the-art methods. At the same time, we retrained the same architecture on CIFARIOO and obtained 86.17% accuracy on the test set, which proves that the architecture obtained by the algorithm has good generalization ability. © 2022 IEEE.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Deep Multi-View Clustering via View-Specific Representation and Global Graph

Deep Multi-View Clustering via View-Specific Representation ...

引用

Artificial Intelligence technology (ACAIT), Asian Conference on

作者： Jingyu Pu Chenhang Cui Yazhou Ren Zichen Wen Tianyi Wu Xiaorong Pu Lifang He School of Computer Science and Engineering University of Electronic Science and Technology of China School of Computer Science and Engineering Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China Department of Computer Science and Engineering Lehigh University Bethlehem USA

ISBN: (数字)9798350359145

ISBN: (纸本)9798350359152

Multi-view clustering can improve clustering performance by leveraging the complementary information from multiview data and has garnered growing interest. Despite significant advancements has been established in this field, most existing methods primarily concentrate on learning the latent representation, neglecting the latent graph that typically reveals relationships between different instances. To address this limitation, this paper presents a new framework of Deep Multi-View Clustering via View-specific Representation and Global Graph (DVGMVC) to address the aforementioned issue. Our method leverages the latent graph of each view to enhance representation quality and incorporates a global graph that combines discriminative information to enhance the performance of deep embedded clustering. Specifically, we employ autoencoders to learn embedded features and construct latent graphs from each view. Subsequently, a weighted two-layer plain residual network refines the latent feature using topological information derived from the latent graphs. As various views adhere to the same foundational cluster structure across multiple views, DVGMVC merges the embedded features of all views to create a unified feature. By leveraging the Graph Convolutional Network (GCN) module, which exploits information from both the latent graphs and the global graph generated from the global feature, a unified embedded feature is generated to capture comprehensive topological information among samples. This unified embedded feature is then employed to generate pseudo labels in a self-supervised manner, guiding unified embedding learning and clustering. In a self-supervised manner, such pseudo-label information guides unified embedding learning and clustering. In-depth experiments on public datasets has proven the state-of-the-art clustering effectiveness of our proposed method.

关键词： Clustering methods Learning (artificial intelligence) Task analysis Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

OCD Diagnosis with Multiple Spatial Similarity-Aware Learning and Diffusion Structure-Aware Graph Convolutional Network

OCD Diagnosis with Multiple Spatial Similarity-Aware Learnin...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Liang, Yu Zhao, Chen Wang, Tianfu Peng, Ziwen Yang, Peng Lei, Baiying Shenzhen University School of Computer Science and Software Shenzhen China Shenzhen University Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging National-Regional Key Technology Engineering Laboratory for Medical Ultrasound School of Biomedical Engineering Shenzhen China South China Normal University Center for the Study of Applied Psychology Guangzhou China

ISBN: (纸本)9798350386226

Obsessive-Compulsive Disorder (OCD) is a hereditary mental illness, and unaffected first-degree relative (UFDR) is also at high risk. This study constructed a framework based on traditional machine learning and deep learning methods to identify OCD and UFDR. Specifically, we propose Multiple Spatial Similarity-Aware Learning (MSSL) models to construct Brain Functional Connectivity Networks (BFCNs) and use Diffusion Structure-Aware Graph Convolutional Network (DSAGCN) for feature learning. Two regularization terms are added to the MSSL model, one is to limit the similarity of functional connectivity of adjacent brain regions, and the other is to constrain the similarity of signals in adjacent brain regions. In this strategy, redundant information can be removed while constructing the interrelationships of various brain regions, which can reflect the interrelationships between brain *** graph diffusion is introduced to the DSAGCN model, which can not only solve the edge noise problem caused by the graph structure constructed by the custom adjacency matrix but also enrich the node features by using the position information of the nodes on the reconstructed graph *** framework has been validated on our dataset collected from local hospitals, and the experimental results show that our proposed method outperforms the state-of-the-art. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：