检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Yuan, Jingyang Mao, Zhengyang Luo, Xiao Ju, Wei Qin, Yifang Zhang, Ming National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University Beijing China Department of Computer Science University of California Los Angeles Los Angeles United States

Graph Neural Networks (GNNs) have garnered considerable interest due to their exceptional performance in a wide range of graph machine learning tasks. Nevertheless, the majority of GNN-based approaches have been examined using well-annotated benchmark datasets, leading to suboptimal performance in real-world graph learning scenarios. To bridge this gap, the present paper investigates the problem of graph transfer learning in the presence of label noise, which transfers knowledge from a noisy source graph to an unlabeled target graph. We introduce a novel technique termed Balance Alignment and Information-aware Examination (ALEX) to address this challenge. ALEX first employs singular value decomposition to generate different views with crucial structural semantics, which help provide robust node representations using graph contrastive learning. To mitigate both label shift and domain shift, we estimate a prior distribution to build subgraphs with balanced label distributions. Building on this foundation, an adversarial domain discriminator is incorporated for the implicit domain alignment of complex multi-modal distributions. Furthermore, we project node representations into a different space, optimizing the mutual information between the projected features and labels. Subsequently, the inconsistency of similarity structures is evaluated to identify noisy samples with potential overfitting. Comprehensive experiments on various benchmark datasets substantiate the outstanding superiority of the proposed ALEX in different settings. Copyright © 2023, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Usability evaluation from client-side traces using heatmap and system usability scale

Usability evaluation from client-side traces using heatmap a...

引用

iNnovative Informatics of Biskra (ISNIB), International Symposium on

作者： Houcem Rezaiguia Abdelhamid Djeffal Computer Science Department LESIA Laboratory Biskra University Algeria

In this project, we investigate the usability of a gamified application developed as part of the e-learning system for the University of Souk Ahras. The online application is integrated within the university web portal, allowing students and lecturers to ask questions and receive answers from other students or lecturers within the university about any modules. This would result in creating a knowledge base of questions and answers for every module that future students could benefit from without the need to ask their lecturers again. We propose the use of heat maps for assessing usability based on client-side traces where a JavaScript app is used to track user behaviors while exploring the university website remotely without accessing or modifying the main university website. The suggested tool is integrated into the webpage by inserting a single line of JavaScript code. After logging and storing the data, a heat map is generated from this data using PHP and JavaScript to assess the most active areas where users payattention when interacting with the website. Furthermore, we created an online implementation of the System Usability Scale questionnaire in order to assess the satisfaction of university participants in using the Question Board platform. The SUS is considered after using the platform. For users who couldn't manage to use the application, we have devised a special questionnaire. Finally, a questionnaire is devised for users who haven't used the application in order to figure out the main reasons for the weak turnout on the platform

关键词： Heating systems Electronic learning Codes Knowledge based systems Relational databases Behavioral sciences Usability

来源：评论

学校读者我要写书评

暂无评论

Unveiling Interactive Narratives: Enhancing MOOCs with Interaction Fiction 4th

Unveiling Interactive Narratives: Enhancing MOOCs with Inte...

引用

4th International Conference on Novel and Intelligent Digital Systems, NiDS 2024

作者： Redjaibia, Abdennour Drissi, Samia Boussaha, Karima Gülseçen, Sevinç Lafifi, Yacine LiM Laboratory University of Mohamed Cherif Messaadia Souk Ahras P.O. Box 1553 Souk Ahras41000 Algeria Department of Computer Science LIAOA Laboratory University of Oum El Bouaghi Oum El Bouaghi Algeria İstanbul University İstanbul Turkey LabSTIC Laboratory University 8 Mai 1945 Guelma P.O. Box 401 Guelma24000 Algeria

ISBN: (纸本)9783031733437

Massive Open Online Courses have become increasingly popular in the last few years. However, they are often characterized by high dropout rates. This paper presents our ongoing research that addresses this issue by proposing a new approach which is integrating Interactive Fiction (IF) into MOOCs with the aid of Artificial Intelligence (AI) especially Large Language Models (LLMs) and Generative AI. The research aims to enhance learner engagement and provide personalized learning experiences by transforming traditional course materials into a series of scenarios. By observing students’ behaviors and interactions within the IF environment, the research seeks to reveal the challenges that learners face during the learning process and provide personalized interventions. Achieving this goal will provide a new dimension to MOOCs, offering enhanced interactivity and individualized support, ultimately contributing to improved learner retention and success. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

RAHNet: Retrieval Augmented Hybrid Network for Long-tailed Graph Classification

arXiv

引用

arXiv 2023年

作者： Mao, Zhengyang Ju, Wei Qin, Yifang Luo, Xiao Zhang, Ming National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University Beijing China Department of Computer Science University of California Los Angeles Los Angeles United States

Graph classification is a crucial task in many real-world multimedia applications, where graphs can represent various multimedia data types such as images, videos, and social networks. Previous efforts have applied graph neural networks (GNNs) in balanced situations where the class distribution is balanced. However, real-world data typically exhibit long-tailed class distributions, resulting in a bias towards the head classes when using GNNs and limited generalization ability over the tail classes. Recent approaches mainly focus on re-balancing different classes during model training, which fails to explicitly introduce new knowledge and sacrifices the performance of the head classes. To address these drawbacks, we propose a novel framework called Retrieval Augmented Hybrid Network (RAHNet) to jointly learn a robust feature extractor and an unbiased classifier in a decoupled manner. In the feature extractor training stage, we develop a graph retrieval module to search for relevant graphs that directly enrich the intra-class diversity for the tail classes. Moreover, we innovatively optimize a category-centered supervised contrastive loss to obtain discriminative representations, which is more suitable for long-tailed scenarios. In the classifier fine-tuning stage, we balance the classifier weights with two weight regularization techniques, i.e., Max-norm and weight decay. Experiments on various popular benchmarks verify the superiority of the proposed method against state-of-the-art approaches. Copyright © 2023, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Technical Report: On the Convergence of Gossip Learning in the Presence of Node Inaccessibility

arXiv

引用

arXiv 2024年

作者： Liu, Tian Cui, Yue Hu, Xueyang Xu, Yecheng Liu, Bo Zhejiang Laboratory Hangzhou China Department of Computer Science and Computer Information Systems Auburn University at Montgomery Montgomery United States Department of Computer Science and Software Engineering Auburn University Auburn United States Institute of Rural Development Zhejiang Academy of Agricultural Sciences Hangzhou China DBAPPSecurity Co. Ltd. Hangzhou China

Gossip learning (GL), as a decentralized alternative to federated learning (FL), is more suitable for resource-constrained wireless networks, such as Flying Ad-Hoc Networks (FANETs) that are formed by unmanned aerial vehicles (UAVs). GL can significantly enhance the efficiency and extend the battery life of UAV networks. Despite the advantages, the performance of GL is strongly affected by data distribution, communication speed, and network connectivity. However, how these factors influence the GL convergence is still unclear. Existing work studied the convergence of GL based on a virtual quantity for the sake of convenience, which failed to reflect the real state of the network when some nodes are inaccessible. In this paper, we formulate and investigate the impact of inaccessible nodes to GL under a dynamic network topology. We first decompose the weight divergence by whether the node is accessible or not. Then, we investigate the GL convergence under the dynamic of node accessibility and theoretically provide how the number of inaccessible nodes, data non-i.i.d.-ness, and duration of inaccessibility affect the convergence. Extensive experiments are carried out in practical settings to comprehensively verify the correctness of our theoretical findings. Copyright © 2024, The Authors. All rights reserved.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Automated Software Vulnerability Detection via Curriculum Learning

Automated Software Vulnerability Detection via Curriculum Le...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Qianjin Du Wei Kun Xiaohui Kuang Xiang Li Gang Zhao Department of Computer Science and Technology Tsinghua University China National Key Laboratory of Science and Technology on Information System Security China

With the development of deep learning, software vulnerability detection methods based on deep learning have achieved great success, which outperform traditional methods in efficiency and precision. At the training stage, all training samples are treated equally and presented in random order. However, in software vulnerability detection tasks, the detection difficulties of different samples vary greatly. Similar to the human learning mechanism following an easy-to-difficult curriculum learning procedure, vulnerability detection models can also benefit from the easy-to-hard curriculums. Motivated by this observation, we introduce curriculum learning for automated software vulnerability detection, which is capable of arranging easy-to-difficult training samples to learn better detection models without any human intervention. Experimental results show that our method achieves obvious performance improvements compared to baseline models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ICM-Assistant: Instruction-tuning Multimodal Large Language Models for Rule-based Explainable Image Content Moderation 39

ICM-Assistant: Instruction-tuning Multimodal Large Language ...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Wu, Mengyang Zhao, Yuzhi Cao, Jialun Xu, Mingjie Jiang, Zhongming Wang, Xuehui Li, Qinbin Hu, Guangneng Qin, Shengchao Fu, Chi-Wing Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong Huawei Hong Kong Research Center Hong Kong Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong Huawei 2012 Laboratories China Artificial Intelligence Institute Shanghai Jiao Tong University China School of Computer Science and Technology Huazhong University of Science and Technology China School of Computer Science and Technology Xidian University China Guangzhou Institute of Technology Xidian University China ICTT and ISN Laboratory Xidian University China

ISBN: (纸本)157735897X

Controversial contents largely inundate the Internet, infringing various cultural norms and child protection standards. Traditional Image Content Moderation (ICM) models fall short in producing precise moderation decisions for diverse standards, while recent multimodal large language models (MLLMs), when adopted to general rule-based ICM, often produce classification and explanation results that are inconsistent with human moderators. Aiming at flexible, explainable, and accurate ICM, we design a novel rule-based dataset generation pipeline, decomposing concise human-defined rules and leveraging well-designed multi-stage prompts to enrich short explicit image annotations. Our ICM-Instruct dataset includes detailed moderation explanation and moderation Q-A pairs. Built upon it, we create our ICM-Assistant model in the framework of rule-based ICM, making it readily applicable in real practice. Our ICM-Assistant model demonstrates exceptional performance and flexibility. Specifically, it significantly outperforms existing approaches on various sources, improving both the moderation classification (36.8% on average) and moderation explanation quality (26.6% on average) consistently over existing MLLMs. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Divergence of Critical Fluctuations on Approaching Catastrophic Phase Inversion in Turbulent Emulsions

引用

Physical Review Letters 2024年第13期133卷 134001页

作者： Lei Yi Ivan Girotto Federico Toschi Chao Sun New Cornerstone Science Laboratory Center for Combustion Energy Key Laboratory for Thermal Science and Power Engineering of Ministry of Education Department of Energy and Power Engineering Department of Physics and Department of Mathematics and Computer Science and J. M. Burgers Centre for Fluid Dynamics 34151 Trieste Italy Via dei Taurini 19 00185 Rome Italy Department of Engineering Mechanics School of Aerospace Engineering

Catastrophic phase inversion, the breakdown of a concentrated emulsion characterized by the most puzzling sudden feature, is crucial in numerous industrial applications. Here we combine well-controlled experiments and fully resolved numerical simulations to study the critical dynamics of catastrophic phase inversion in oil-water emulsions under turbulent flow as the phase-inversion volume fraction is approached. We reveal that the phase inversion is characterized by the critical power-law divergence of fluctuations in the global drag force. We determine the enhanced dynamical heterogeneity in the local droplet structures at approaching the phase inversion, and tightly connect it to the diverging drag fluctuations. Moreover, we show that near to the critical point the phase inversion is triggered as a stochastic process by large fluctuations at both large and small scales. Our findings pave the way to modeling the phase inversion process as an out-of-equilibrium critical-like phenomena.

关键词： Drop & bubble phenomena Granular packing Jamming Turbulent multiphase flows

来源：评论

学校读者我要写书评

暂无评论

Full VNIR-SWIR hyperspectral imaging workflow for the monitoring of archaeological textiles

Full VNIR-SWIR hyperspectral imaging workflow for the monito...

引用

2023 IS and T Archiving Conference, ARCHIVING 2023

作者： Grillini, Federico Thomas, Jean-Baptiste George, Sony Department of Computer Science Norwegian University of Science and Technology Gjøvik Norway ImViA Laboratory University of Burgundy Dijon France

ISBN: (纸本)9781713890997

A practical workflow to capture and process hyperspectral images in combined VNIR-SWIR ranges is presented and discussed. The pipeline demonstration is intended to increase the visibility of the possibilities that advanced hyperspectral imaging techniques can bring to the study of archaeological textiles. Emphasis is placed on the fusion of data from two hyperspec-tral devices. Every aspect of the pipeline is analyzed, from the practical and optimal implementation of the imaging setup to the choices and decisions that can be made during the data processing steps. The workflow is demonstrated on an archaeological textile belonging to the Paracas Culture (Peru, 200 BC -100 AD ca.) and displays an example in which an inappropriate selection of the processing steps can lead to a misinterpretation of the hyperspectral data. © 2023 Society for Imaging science and Technology.

关键词： Fabrics

来源：评论

学校读者我要写书评

暂无评论

Fast PCB Stack-Up Optimization Using Integer Programming

IEEE Transactions on Signal and Power Integrity

引用

IEEE Transactions on Signal and Power Integrity 2023年 2卷 32-42页

作者： He, Jiayi Zhang, Ling Kiguradze, Zurab Chada, Arun Klivans, Adam Mutnury, Bhyrav Li, Er-Ping Fan, Jun Missouri University of Science and Technology Electromagnetic Compatibility Laboratory RollaMO65409 United States Zhejiang University College of Information Science and Electronic Engineering Hangzhou310027 China Enterprise Product Group in Dell Round RockTX78682 United States University of Texas at Austin Department of Computer Science AustinTX78712-0803 United States College of Information Science and Electronic Engineering Zhejiang University Hangzhou310027 China

This article presents a flexible and efficient methodology to optimize stack-up for multilayer printed circuit boards (PCBs) with enormous search space and various design constraints. PCB stack-up optimization is crucial in high-speed system design to achieve the desired electrical performance while reducing system costs. The stack-up optimization process is labor-intensive and time-consuming for a large number of layers. Moreover, after the optimization process, the electrical performance of a real design, such as the impedance and loss, may deviate from the target design due to manufacturing variations. Estimating the worst cases due to the manufacturing variations, referred to as 'corner cases' in this article, is essential for a confident PCB design but challenging since the number of related parameters is large. In this article, PCB stack-up optimization and corner-case searching are addressed and greatly accelerated using the integer programming technique. All constraints are converted to mathematical equalities and inequalities that can be solved rapidly by an integer programming solver to obtain feasible stack-up solutions. After the cross sections of the transmission lines are optimized based on the stack-up design to achieve a target electrical performance, the upper and lower bound of impedance and loss are acquired using integer programming when the design parameters vary in a particular range. The proposed method is verified using multilayer PCB designs with practical constraints and demonstrates its effectiveness and high efficiency. © 2022 IEEE.

关键词： Printed circuit boards

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：