检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Guo, Xianjie Yu, Kui Cui, Lizhen Yu, Han Li, Xiaoxiao School of Computer Science and Information Engineering Hefei University of Technology China Key Laboratory of Knowledge Engineering with Big Data of Ministry of Education China School of Software Shandong University China College of Computing and Data Science Nanyang Technological University Singapore Department of Electrical and Computer Engineering The University of British Columbia Canada Vector Institute Canada

ISBN: (纸本)157735897X

Federated feature selection (FFS) is a promising field for selecting informative features while preserving data privacy in federated learning (FL) settings. Existing FFS methods focus on capturing the correlations between features and labels. They struggle to achieve satisfactory performance in the face of data distribution heterogeneity among FL clients, and cannot address the out-of-distribution (OOD) problem that arises when a significant portion of clients do not actively participate in FL training. To address these limitations, we propose Federated Causally Invariant Feature Learning (FedCIFL), a novel approach for learning causally invariant features in a privacy-preserving manner. We design a sample reweighting strategy to eliminate spurious correlations introduced by selection bias and iteratively estimate the federated causal effect between each feature and the labels (with the remaining features initially treated as confounders). By iteratively refining the confounding feature set to identify the true confounders, FedCIFL mitigates the impact of limited local data on the accuracy of federated causal effect estimation. Theoretical analysis proves the correctness of FedCIFL under reasonable assumptions. Extensive experiments on synthetic and real-world datasets demonstrate the superiority of FedCIFL against eight state-of-the-art baselines, beating the best-performing approach by 3.19%, 9.07% and 2.65% in terms of average test Accuracy, RMSE and F1 score, respectively. It is a first-of-its-kind FFS approach capable of handling Non-IID and OOD data simultaneously. The source code is available at https://***/Xianjie-Guo/FedCIFL. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Personalized Federated Learning with Collaborative Aggregation Networks for Multi-Site Brain Disorder Diagnosis 4

Personalized Federated Learning with Collaborative Aggregati...

引用

4th International Conference on Industrial Automation, Robotics and Control engineering, IARCE 2024

作者： Si, Qian Li, Yang School of Cyber Science and Technology Beihang University Beijing China Department of Automation Science and Electrical Engineering The Beijing Advanced Innovation Center for Big Data and Brain Computing State Key Laboratory of Virtual Reality Technology and Systems Beijing China Advanced Institute of Information Technology Peking University Beijing China Beihang University Beijing China

ISBN: (纸本)9798350380323

In multi-site brain disease diagnosis studies, traditional centralized training methods necessitate sharing medical data, posing significant privacy risks. Federated learning (FL) offers a privacy-preserving solution by enabling global model training through aggregating locally trained models from multiple data centers without sharing raw data. However, current FL approaches rely on a server-based network topology, where central server failure disrupts training. Additionally, data heterogeneity across sites often slows convergence and reduces accuracy. To overcome these issues, we introduce a decentralized personalized federated learning collaborative aggregation network (pFedCAN). This framework has two core components: (1) separating local models into shared and personalized layers, and (2) forming a collaborative aggregation network via similarity detection in the shared layers. Specifically, each center trains its local model, then separates it into shared and personalized layers. The shared layer is exchanged with other centers, while the personalized layer remains local. data centers analyze similarities in received shared layers to build a collaborative network, where shared layers from similar centers are aggregated to refine the model. This approach flexibly adapts to varying levels of data heterogeneity, enhancing model training efficiency. Validation on public datasets, ABIDE I and ADHD, shows that the proposed method outperforms current leading techniques. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Integrating Attribute information into Relational data: A Latent Structure Approach

Integrating Attribute Information into Relational Data: A La...

引用

Consumer Technology (ICCT-Pacific), International Conference on

作者： Kenta Mikawa Manabu Kobayashi Tomoyuki Sasaki Akiko Manada Department of Information Systems Tokyo City University Kanagawa Japan Center for Data Science Waseda University Tokyo Japan Department of Computer Science Tokyo City University Tokyo Japan Department of Electrical Electronics and Information Engineering Nagaoka University of Technology Niigata Japan

ISBN: (数字)9798331504120

ISBN: (纸本)9798331504137

Various analytical methods for diverse data have been proposed with the evolution of information technology. This study focuses on relational data, comprising information on relationships between objects. Although several studies have focused on the relationships between objects, only a few have utilized attribute information. The conventional study proposed analyzing relational data with attribute information using a latent structure model. However, the model assumes that both attribute information and relational data are generated by a single latent variable. Furthermore the model does not effectively handle missing values in relational data. Therefore, in this study, we modify the conventional latent structure analysis model for relational data with attribute information, which assumes that latent variables have various types of structures, and derived a method for estimating its parameters. In addition, we apply the proposed model to cases when the relational data contain missing values. The effectiveness of our proposed method is demonstrated through simulation experiments on synthetic dataset.

关键词： Analytical models data models information technology Synthetic data

来源：评论

学校读者我要写书评

暂无评论

Adversarial Attacks on Featureless Deep Learning Malicious URLs Detection

引用

Computers, Materials & Continua 2021年第7期68卷 921-939页

作者： Bader Rasheed Adil Khan S.M.Ahsan Kazmi Rasheed Hussain Md.Jalil Piran Doug Young Suh Institute of Data Science and Articial Intelligence Innopolis UniversityInnopolis420500Russia Institute of Information Security and Cyberphysical Systems Innopolis UniversityInnopolis420500Russia Department of Computer Science and Engineering Sejong UniversitySeoulKorea Department of Electronics Engineering Kyung Hee UniversityYonginKorea

Detecting malicious Uniform Resource Locators(URLs)is crucially important to prevent attackers from committing *** researches have investigated the role of machine learning(ML)models to detect malicious *** using ML algorithms,rst,the features of URLs are extracted,and then different ML models are *** limitation of this approach is that it requires manual feature engineering and it does not consider the sequential patterns in the ***,deep learning(DL)models are used to solve these issues since they are able to perform featureless ***,DL models give better accuracy and generalization to newly designed URLs;however,the results of our study show that these models,such as any other DL models,can be susceptible to adversarial *** this paper,we examine the robustness of these models and demonstrate the importance of considering this susceptibility before applying such detection systems in real-world *** propose and demonstrate a black-box attack based on scoring functions with greedy search for the minimum number of perturbations leading to a *** attack is examined against different types of convolutional neural networks(CNN)-based URL classiers and it causes a tangible decrease in the accuracy with more than 56%reduction in the accuracy of the best classier(among the selected classiers for this work).Moreover,adversarial training shows promising results in reducing the inuence of the attack on the robustness of the model to less than 7%on average.

关键词： Malicious URLs detection deep learning adversarial attack web security

来源：评论

学校读者我要写书评

暂无评论

Softsign Gaussian Deep Belief Neural Network (SGDBNN): A Novel Technique for Improving Pedestrian Detection Accuracy and Processing Efficiency 7

Softsign Gaussian Deep Belief Neural Network (SGDBNN): A Nov...

引用

7th International Conference on Image information Processing, ICIIP 2023

作者： Bhavani, N.P.G. Sudha, R.V. Priya, K.Shanmuga Arun Raj, S.R. Yadav, Vinay Kumar Balaram, Allam Saveetha School of Engineering Saveetha Institute of Medical And Technical Sciences Department of Electronics And Communication Engineering Tamilnadu India Knowledge Institute of Technology Department of Computer Science And Business Systems Selliampalayam Tamil Nadu Kakapalayam India KSR Institute For Engineering And Technology Department of Information Technology TamilNadu Tiruchengode637215 India U.B.D.T College of Engineering Department of Electronics & Communication Engineering Karnataka Davanagere577004 India Glocal University Glocal School of Science And Technology Uttar Pradesh India Mlr Institute of Technology Department of Computer Science And Engineering Hyderabad500043 India

ISBN: (纸本)9798350371406

Pedestrian detection is a critical task in computer vision with the aim of accurately identifying pedestrians in images or video frames. In this work, we tackle the issue of enhancing pedestrian detection precision while reducing processing time. To achieve this, we propose a novel technique called Softsign Gaussian Deep Belief Neural Network (SGDBNN) for pedestrian detection. The proposed SGDBNN technique leverages the power of recurrent deep neural networks to handle the complexities of pedestrian detection. The input layer receives images as input, which are then processed in the hidden layer. Here, a Gaussian activation function is employed to extract relevant features, effectively detecting pedestrian objects in the images. The incorporation of deep learning and recurrent behavior in the neural networks results in reduced incorrect classifications, leading to a lower false positive rate and improved pedestrian detection accuracy. The softsign activation function at the hidden layer allows the proposed technique to efficiently compare extracted features with pre-stored templates, enabling effective pedestrian identification in the images. By effectively utilizing the recurrent behavior and advanced activation functions, the SGDBNN demonstrates superior performance compared to existing methods, such as Deep Convolutional Neural Networks and Extended Deep Model. Experimental evaluations illustrate the effectiveness of the proposed SGDBNN technique. It achieves a significant increase in Positive Detection Accuracy (PDA) of up to 20% and 16%, while simultaneously reducing the False Positive Rate (FPR) by up to 18% and 3% compared to existing Deep Convolutional Neural Networks and Extended Deep Model, respectively. Additionally, the proposed SGDBNN reduces training time by up to 30% and 12% and decreases memory requirements by up to 47% and 30% when compared to existing Deep Convolutional Neural Networks and Extended Deep Model, respectively. Likewise, the SGDBNN tec

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Testing of Emerging Wireless Sensor Networks Using Radar Signals With Machine Learning Algorithms

IEEE Journal of Selected Areas in Sensors

引用

IEEE Journal of Selected Areas in Sensors 2024年 1卷 49-59页

作者： Shitharth Selvarajan Hariprasath Manoharan Adil O. Khadidos Alaa O. Khadidos School of Built Environment Engineering and Computing Leeds Beckett University Leeds U.K. Department of Electronics and Communication Engineering Panimalar Engineering College Chennai India Department of Information Technology Faculty of Computing and Information Technology King Abdulaziz University Jeddah Saudi Arabia Department of Information Systems Faculty of Computing and Information Technology King Abdulaziz University Jeddah Saudi Arabia Center of Research Excellence in Artificial Intelligence and Data Science King Abdulaziz University Jeddah Saudi Arabia

In this article, machine learning methods are used to assess how well wireless sensor networks transmit and receive radar signals. Measurements are done with labeled and unlabeled datasets where output functions are modified in relation to transmitted input in order to test the transceiver of radar signals. The main contribution in the proposed method is to focus on the possibility of choosing a free space model that transmits the radar signals in wireless sensor networks without any interruptions. Hence, for such type of transmissions, reference time period is selected in order to perform radar signal classification, and at the same time, separation of unnecessary interruptions is reduced using clustering procedures. Since the radar signals can be monitored with automatic transmission techniques, the outcomes are combined with supervised, unsupervised, and reinforcement learning models to increase the effect of transmissions. Therefore, the objective functions are designed with three scenarios where reinforcement learning proves to provide adequate connections for radar signals to all wireless sensor networks at reduced error of 0.3%. In addition, with reinforcement learning, the distance of radar signal transmission is maximized to a level greater than 75% at minimized noise ratio of 0.8%.

关键词： Radar Wireless sensor networks Sensors Spaceborne radar Security Radar measurements Mathematical models

来源：评论

学校读者我要写书评

暂无评论

Enhancement of Security in IoT using Secure Hash Cryptography Algorithm with Double Encryption (SHCA-DE) 9

Enhancement of Security in IoT using Secure Hash Cryptograph...

引用

9th International Conference on Communication and Electronics systems, ICCES 2024

作者： Uganya, G. Rajalakshmi, D. Vangapandu, Ramarao Dinesh, M. Giri, Jayant Odeh, Mahmoud Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Department of Ece Chennai India Srm Institute of Science and Technology Department of Data Science and Business Systems Kattankulathur Chennai India Vignan's Institute of Engineering for Women Department of Computer Science and Engineering Vishakhapatnam India Saveetha University Department of Vlsi Microelectronics Saveetha School of Engineering Chennai India Yeshwantrao Chavan College of Engineering Department of Mechanical Engineering Nagpur India Lovely Professional University Division of Research and Development Phagwara India Zarqa University Faculty of Information Technology Department of Cyber Security Zarqa Jordan University of Business and Technology Jeddah Saudi Arabia

ISBN: (纸本)9798350377972

The 'Internet of things' is the developing technology to transfer the world into smart innovation. 'IoT' security is the most problematic issue in recent applications due to its properties including 'centralized architecture', 'limited capacity devices', and 'ambiguity'. So, blockchain with 'IoT' applications provides secure data communication compared to traditional 'IoT' applications. Blockchain is a recent growing technology that has features including 'decentralized structure', 'transparency', 'immutable records', 'increased capacity', and 'quick backtracks'. The 'cryptography hash algorithm' is very important to provide secure data transmission. It converted the different size input to fixed-size unalterable output to enhance the security. But still, it needs efficient algorithm to secure data transmission with safeguard of private key. The objective of this paper is to propose a secure hash cryptography algorithm with double encryption (SHCA-DE) to protect the private key message. Then, this proposed algorithm has been analyzed with MD5, SHA-256, and SHA3-512. From this analysis, the SHA3-512 performs well with proposed algorithm. © 2024 IEEE.

关键词： Encryption algorithms

来源：评论

学校读者我要写书评

暂无评论

IKGN: Intention-aware knowledge Graph Network for POI Recommendation

IKGN: Intention-aware Knowledge Graph Network for POI Recomm...

引用

IEEE International Conference on data Mining (ICDM)

作者： Xiaoyu Zhu Chenyang Bu Bingbing Dong Shengwei Ji Yi He Xindong Wu Key Laboratory of Knowledge Engineering with Big Data (the Ministry of Education of China) Hefei University of Technology School of Computer Science and Information Engineering Hefei University of Technology China School of Artificial Intelligence and Big Data Hefei University China Department of Computer Science Old Dominion University USA Research Center for Knowledge Engineering Zhejiang Lab China

Point-of-Interest (POI) recommendation, pivotal for guiding users to their next interested locale, grapples with the persistent challenge of data sparsity. Whereas knowledge graphs (KGs) have emerged as a favored tool to mitigate the issue, existing KG-based methods tend to overlook two crucial elements: the intention steering users’ location choices and the high-order topological structure within the KG. In this paper, we craft an Intention-aware knowledge Graph (IKG) that harmonizes users’ visit histories, movement trajectories, and location categories to model user intentions. Building upon IKG, our novel Intention-aware knowledge Graph Network (IKGN) delves deeper into the POI recommendation by weighing and propagating node embeddings through an attention mechanism, capturing the unique locational intent of each user. A sequential model like GRU is then employed to ensure a comprehensive representation of users’ short- and long-term location preferences. An empirical study on two real-world datasets validates the effectiveness of our proposed IKGN, with it markedly outshining seven benchmark rival models in both Recall and NDCG metrics. The code of IKGN is available at https://***/Jungle123456/IKGN.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Co-training partial domain adaptation networks for industrial Fault Diagnosis

arXiv

引用

arXiv 2024年

作者： Chen, Gecheng Luo, Chengwen Li, Jianqiang Chen, Xinkai The National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China The Department of Electronic Information Systems Shibaura Institute of Technology Saitama337-8570 Japan

The partial domain adaptation (PDA) challenge is a prevalent issue in industrial fault diagnosis. Current PDA approaches primarily rely on adversarial learning for domain adaptation and use reweighting strategies to exclude source samples deemed outliers. However, the transferability of features diminishes from general feature extraction layers to higher task-specific layers in adversarial learning-based adaptation modules, leading to significant negative transfer in PDA settings. We term this issue the adaptation-discrimination paradox (ADP). Furthermore, reweighting strategies often suffer from unreliable pseudo-labels, compromising their effectiveness. Drawing inspiration from traditional classification settings where such partial challenge is not a concern, we propose a novel PDA framework called Interactive Residual Domain Adaptation Networks (IRDAN), which introduces domain-wise models for each domain to provide a new perspective for the PDA challenge. Each domain-wise model is equipped with a residual domain adaptation (RDA) block to mitigate the ADP problem. Additionally, we introduce a confident information flow via an interactive learning strategy, training the modules of IRDAN sequentially to avoid cross-interference. We also establish a reliable stopping criterion for selecting the best-performing model, ensuring practical usability in real-world applications. Experiments have demonstrated the superior performance of the proposed IRDAN. Copyright © 2024, The Authors. All rights reserved.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

The Smart Application of data Mining in the Detection of Fraudulent Transactions 3

The Smart Application of Data Mining in the Detection of Fra...

引用

3rd International Conference on Smart Generation Computing, Communication and Networking, SMART GENCON 2023

作者： Bodhankar, Aniruddha Aswal, Upendra Singh Arivazhagn, D. Pund, Sachin S. Kulkarni, Nandini Chaturvedi, Sudhir Kumar Dr. Ambedkar Institute of Management Studies and Research- Nagpur Department of Data Science & Information Systems Nagpur India Shri Ramdeobaba College of Engineering and Management Department of Industrial Engineering Maharashtra Nagpur India Graphic Era Deemed to Be University Department of Computer Science & Engineering Uttarakhand Dehradun India Symbiosis International University Department of Symbiosis School of Planning Architecture and Design Nagpur Nagpur India Business School Academy of Maritime Education and Training Deemed to Be University Department of Amet Chennai India Upes Department of School of Engineering Dehradun India

ISBN: (纸本)9798350319125

Records mining is a powerful analytical device used to find out styles, correlations, and traits in large facts sets. Smart application of data mining methods within the detection of fraudulent transactions has finished an extensive quantity of success. It normally entails statistical algorithms, facts evaluation, and machine getting to know techniques to stumble on suspicious sports associated with fraud. Using statistics mining techniques, it's far possible to music the behaviour of clients, discover anomalous transaction styles and developments in huge datasets, come across frauds and generate signals on suspicious activities. It also allows to identify which forms of merchandise are frequently centered, which price methods are used, which places are problem to danger, and which traders have capacity credit card fraud. The utility of facts mining within the detection of fraudulent sports entails principal techniques. The first method makes use of conventional techniques used by banks, credit groups and other retailers, which includes account reconciliation, annualized billing and complex question processing. Those methods are used to generate reviews that are used by investigators to perceive probably suspicious activities. The second technique entails the usage of superior analytical techniques such as neural networks, decision timber, and clustering algorithms. These strategies have the functionality to stumble on facts irregularities in several conditions that were formerly unseen or unknown. © 2023 IEEE.

关键词： developments entails reconciliation statistics transaction

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：