检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Wu, Bingjie Yu, Zitong Xie, Yiping Liu, Wei Luo, Chaoqi Liu, Yong Goh, Rick Siow Mong Singapore School of Computing and Information Technology Great Bay University Dongguan523000 China Computer Vision Institute School of Computer Science & Software Engineering Shenzhen University Shenzhen518060 China

Remote Photoplethysmography (rPPG) is a promising technique to monitor physiological signals such as heart rate from facial videos. However, the labeled facial videos in this research are challenging to collect. Current rPPG research is mainly based on several small public datasets collected in simple environments, which limits the generalization and scale of the AI models. Semi-supervised methods that leverage a small amount of labeled data and abundant unlabeled data can fill this gap for rPPG learning. In this study, a novel semi-supervised learning method named Semi-rPPG that combines curriculum pseudo-labeling and consistency regularization is proposed to extract intrinsic physiological features from unlabelled data without impairing the model from noises. Specifically, a curriculum pseudo-labeling strategy with signal-to-noise ratio (SNR) criteria is proposed to annotate the unlabelled data while adaptively filtering out the low-quality unlabelled data. Besides, a novel consistency regularization term for quasi-periodic signals is proposed through weak and strong augmented clips. To benefit the research on semi-supervised rPPG measurement, we establish a novel semi-supervised benchmark for rPPG learning through intra-dataset and cross-dataset evaluation on four public datasets. The proposed Semi-rPPG method achieves the best results compared with three classical semi-supervised methods under different protocols. Ablation studies are conducted to prove the effectiveness of the proposed methods. Copyright © 2025, The Authors. All rights reserved.

关键词： Curricula

来源：评论

学校读者我要写书评

暂无评论

Optimizing Secure Multi-User ISAC Systems With STAR-RIS: A Deep Reinforcement Learning Approach for 6G Networks

引用

IEEE Access 2025年 13卷 31472-31484页

作者： Kamal, Mian Muhammad Zain Ul Abideen, Syed Al-Khasawneh, M.A. Alabrah, Amerah Sohail Ahmed Larik, Raja Irfan Marwat, Muhammad Southeast University School of Electronic Science and Engineering Jiangning Jiangsu Nanjing211189 China Qingdao University College of Computer Science and Technology Qingdao266071 China Al-Ahliyya Amman University Hourani Center for Applied Scientific Research Amman19111 Jordan Skyline University College School of Computing University City Sharjah Sharjah United Arab Emirates King Saud University College of Computer and Information Science Department of Information Systems Riyadh11543 Saudi Arabia Ilma University Department of Computer Science Sindh Karachi75190 Pakistan University of Science and Technology Bannu Department of Software Engineering Bannu28100 Pakistan

The rapid evolution of wireless communication technologies and the increasing demand for multi-functional systems have led to the emergence of integrated sensing and communication (ISAC) as a key enabler for future 6G networks. Simultaneously transmitting and reflecting reconfigurable intelligent surfaces (STAR-RISs) have recently garnered significant attention for their ability to enhance signal coverage and improve system efficiency. This paper investigates a STAR-RIS-assisted ISAC system designed to secure communication for multiple legitimate users (LUs) while safeguarding against multiple eavesdroppers (Eves). By jointly optimizing the base station (BS) transmit beamforming, STAR-RIS transmission and reflection coefficients, and receive filters, the proposed framework aims to maximize the long-term average secrecy rate for all LUs. Constraints are imposed to ensure minimum echo signal-to-noise ratios (SNRs) for sensing and meet the achievable rate requirements of LUs. To address the inherent complexity of this non-convex problem, two deep reinforcement learning (DRL) algorithms are proposed. Numerical results demonstrate that the proposed system achieves significant improvements in secrecy rate compared to conventional RIS setups. This work provides a scalable and efficient approach for secure multi-user ISAC systems, making it highly relevant for future 6G networks, smart cities, and IoT applications. © 2025 The Authors.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Instructors' Perspectives on Capstone Courses in computing Fields: A Mixed-Methods Study 2024

Instructors' Perspectives on Capstone Courses in Computing F...

引用

29th Working Group Reports on Innovation and Technology in Computer Science Education, ITiCSE WGR 2024

作者： Hooshangi, Sara Shakil, Asma Dasgupta, Subhasish Davis, Karen C.C. Farghally, Mohammed Fitzpatrick, KellyAnn Gutica, Mirela Hardt, Ryan Riddle, Steve Seyam, Mohammed Department of Computer Science Virginia Tech BlacksburgVA United States School of Computer Science University of Auckland Auckland New Zealand Info Systems and Tech Mgmt Dept George Washington University WashingtonDC United States Computer Science and Software Engineering Dept. Miami University OxfordOH United States School of Computing Instruction Georgia Institute of Technology AtlantaGA United States School of Computing & Academic Studies British Columbia Institute of Technology BurnabyBC Canada Computer & Information Science University of St. Thomas Saint PaulVA United States School of Computing Newcastle University Newcastle upon Tyne United Kingdom

ISBN: (纸本)9798400712081

Team-based capstone courses are integral to many undergraduate and postgraduate degree programs in the computing field. They are designed to help students gain hands-on experience and practice professional skills such as communication, teamwork, and self-reflection as they transition into the real world. Prior research on capstone courses has focused primarily on the experiences of students. The perspectives of instructors who teach capstone courses have not been explored comprehensively. However, an instructor's experience, motivation, and expectancy can have a significant impact on the quality of a capstone course. In this working group, we used a mixed methods approach to understand the experiences of capstone instructors. Issues such as class size, industry partnerships, managing student conflicts, and factors influencing instructor motivation were examined using a quantitative survey and semi-structured interviews with capstone teaching staff from multiple institutions across different continents. Our findings show that there are more similarities than differences across various capstone course structures. Similarities include team size, team formation methodologies, duration of the capstone course, and project sourcing. Differences in capstone courses include class sizes and institutional support. Some instructors felt that capstone courses require more time and effort than regular lecture-based courses. These instructors cited that the additional time and effort is related to class size and liaising with external stakeholders, including industry partners. Some instructors felt that their contributions were not recognized enough by the leadership at their institutions. Others acknowledged institutional support and the value that the capstone brought to their department. Overall, we found that capstone instructors were highly intrinsically motivated and enjoyed teaching the capstone course. Most of them agree that the course contributes to their professional de

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Runtime Backdoor Detection for Federated Learning via Representational Dissimilarity Analysis

引用

IEEE Transactions on Dependable and Secure computing 2025年

作者： Zhang, Xiyue Xue, Xiaoyong Du, Xiaoning Xie, Xiaofei Liu, Yang Sun, Meng University of Oxford Department of Computer Science United Kingdom Peking University Department of Information and Computing Science School of Mathematical Science China Monash University Department of Software Systems and Cybersecurity Australia Singapore Management University School of Computing and Information Systems Singapore Nanyang Technological University School of Computer Science and Engineering Singapore

Federated learning (FL), as a powerful learning paradigm, trains a shared model by aggregating model updates from distributed clients. However, the decoupling of model learning from local data makes FL highly vulnerable to backdoor attacks, where a single compromised client can poison the shared model. While recent progress has been made in backdoor detection, existing methods face challenges with detection accuracy and runtime effectiveness, particularly when dealing with complex model architectures. In this work, we propose a novel approach to detecting malicious clients in an accurate, stable, and efficient manner. Our method utilizes a sampling-based network representation method to quantify dissimilarities between clients, identifying model deviations caused by backdoor injections. We also propose an iterative algorithm to progressively detect and exclude malicious clients as outliers based on these dissimilarity measurements. Evaluations across a range of benchmark tasks demonstrate that our approach outperforms state-of-the-art methods in detection accuracy and defense effectiveness. When deployed for runtime protection, our approach effectively eliminates backdoor injections with marginal overheads. © 2025 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Reliable Service Recommendation: A Multi-modal Adversarial Method for Personalized Recommendation under Uncertain Missing Modalities

引用

IEEE Transactions on Services computing 2025年

作者： Chen, Junyang Yang, Ruohan Guo, Jingcai Wang, Huan Wu, Kaishun Zhang, Liangjie Shenzhen University College of Computer Science and Software Engineering China Huazhong Agricultural University College of Informatics Wuhan China The Hong Kong Polytechnic University Department of Computing Hong Kong DSA Thrust and IoT Thrust Center China

Personalized recommendation is of paramount importance in online content platforms like Kuai and Tencent. To ensure accurate recommendations, it is crucial to consider multi-modal information in both items and user-user/item interactions. While existing works on multimedia recommendation have made strides in leveraging multi-modal contents to enrich item representations, many of them overlook the practical scenario of multiple modality missing. As a result, the performance of recommendation systems can be significantly compromised in such cases. In this paper, we introduce a novel multi-modal adversarial method called MMAM, which aims to provide reliable personalized recommendation services even in the presence of uncertain missing modalities. The core idea behind MMAM is to design a generator that can effectively encode both user-user/item interactions and multi-modal contents, taking into account various missing cases. The generator is trained to learn transferable features from different combinations of missing modalities in order to deceive a discriminative classifier. Additionally, we propose a modal discriminator that can classify the missing cases of multi-modalities, further enhancing the capability of the model. Moreover, a well-equipped predictor utilizes the transferable features to predict potential user interests. To improve the prediction accuracy, we design a type discriminator that enhances the classification of link types. By employing a mini-max game between the generator and the discriminators, MMAM successfully obtains transferable features that encompass multi-modal contents, even when facing uncertain missing modalities. We conduct extensive experiments on industrial datasets, including Kuai and Tencent. Comparing with state-of-the-art approaches, MMAM achieves improvements in personalized recommendation tasks under uncertain missing modalities. MMAM holds promise for enhancing multi-modal personalized recommendations in real-world applications

关键词： Discriminators

来源：评论

学校读者我要写书评

暂无评论

Evaluation of machine learning-based regression techniques for prediction of diabetes levels fluctuations

引用

Heliyon 2025年第1期11卷 e41199页

作者： Alkalifah, Badriah Shaheen, Muhammad Tariq Alotibi, Johrah Alsubait, Tahani Alhakami, Hosam Department of Computer Science and Artificial Intelligence College of Computing Umm Al-Qura University Makkah Saudi Arabia School of Electrical Engineering and Computer Science NUST Islamabad Pakistan Department of Software Engineering College of Computing Umm Al-Qura University Makkah Saudi Arabia

Middle-Aged and Elderly people today face a variety of health problems as a result of their modern lifestyle, which includes increased work stress, less physical activity, and altered food habits. Because of Complications arising, diabetes has become one of the most frequent, severe, and fatal illnesses around the world. Therefore, inaccurate measurements of blood glucose levels can seriously damage vital organs. Several strategies for long-term glucose prediction have been proposed in the literature. Unfortunately, these methods require the patient to identify their daily activities, which can be error-prone, such as meal intake, insulin injection, and emotional aspects. This paper suggests using continuous glucose monitoring (CGM) of 14733 patients, with three assistance factors to predict blood glucose levels independently of other parameters, hence reducing the burden on the patients. To support this an Artificial Neural Network (ANN), Binary Decision Tree (BDT), Linear Regression (LR), Boosting Regression Tree Ensemble (BSTE), Linear Regression with Stochastic Gradient Descent (LRSGD), Stepwise (SW), Support Vector Machine (SVM), and Gaussian process regression (GPR) were investigated. The result indicated that The highest classification accuracy of (92.58%) has been achieved by BDT followed by BSTE (92.04%) and GPR (88.59%). The obtained average of root means square error (MSE) was 1.64, 1.67, 1.69, mg/dL for prediction horizon (PH) respectively to GPR, BSTE, and ANN. © 2024 The Author(s)

关键词： Continuous glucose monitoring (CGM) Diabetic Hyperglycemia Hypoglycemia Machine learning (ML) Regression

来源：评论

学校读者我要写书评

暂无评论

Optimal Additive Noise Mechanisms for Differential Privacy

arXiv

引用

arXiv 2025年

作者： Gilani, Atefeh Gomez, Juan Felipe Asoodeh, Shahab Calmon, Flavio P. Kosut, Oliver Sankar, Lalitha School of Electrical Computer and Energy Engineering Arizona State University United States Department of Physics Harvard University United States Department of Computing and Software McMaster University Canada School of Engineering and Applied Sciences Harvard University United States

We propose a unified optimization framework for designing continuous and discrete noise distributions that ensure differential privacy (DP) by minimizing Rényi DP, a variant of DP, under a cost constraint. Rényi DP has the advantage that by considering different values of the Rényi parameter α, we can tailor our optimization for any number of compositions. To solve the optimization problem, we reduce it to a finite-dimensional convex formulation and perform preconditioned gradient descent. The resulting noise distributions are then compared to their Gaussian and Laplace counterparts. Numerical results demonstrate that our optimized distributions are consistently better, with significant improvements in (Ε, δ)-DP guarantees in the moderate composition regimes, compared to Gaussian and Laplace distributions with the same variance. © 2025, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Verilua: An Open Source Versatile Framework for Efficient Hardware Verification and Analysis Using LuaJIT

Verilua: An Open Source Versatile Framework for Efficient Ha...

引用

Design, Automation and Test in Europe Conference and Exhibition

作者： Ye Cai Chuyu Zheng Wei He Dan Tang College of Computer Science and Software Engineering Shenzhen University Shenzhen China Beijing Institute of Open Source Chip Beijing China Institute of Computing Technology Chinese Academy of Sciences (lCT) Beijing China

ISBN: (数字)9783982674100

ISBN: (纸本)9798331534646

The growing complexity of hardware verification highlights limitations in existing frameworks, particularly regarding flexibility and reusability. Current methodologies often require multiple specialized environments for functional verification, waveform analysis, and simulation, leading to toolchain fragmentation and inefficient code reuse. This paper presents Verilua, a unified framework leveraging LuaJIT and the Verilog Procedural Interface (VPI), which integrates three core functional-ities: Lua-based functional verification, a scripting engine for RTL simulation, and waveform analysis. By enabling complete code reuse through a unified Lua codebase, the framework achieves a 12x speedup in RTL simulation compared to cocotb and a 70x improvement in waveform analysis over state-of-the-art solutions. Through consolidating verification tasks into a single platform, Verilua enhances efficiency while reducing tool fragmentation and learning overhead, addressing critical challenges in modern hardware design.

关键词： Codes Europe Hardware Complexity theory Signal analysis Hardware design languages Engines

来源：评论

学校读者我要写书评

暂无评论

Fused Fuzzy Deep Learning and Information Steganography for Privacy Preservation in Medical Consumer Electronics

引用

IEEE Transactions on Consumer Electronics 2025年

作者： Chen, Jingxue Zhao, Pengbiao Deng, Erqiang Khan, Fazlullah Alturki, Ryan Alshawi, Bandar Xiong, Hu School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu610054 China the School of Computer Science Faculty of Science and Engineering University of Nottingham Ningbo China Zhejiang Ningbo315104 China the KKR and KSR institute of technology and sciences Andhra Pradesh Guntur India the Division of Research and Development Lovely Professional University Phagwara India the University Centre of Research and Development Chandigarh University Mohali India the Department of Computer and Network Engineering College of Computing Umm Al-Qura University Makkah Saudi Arabia the Department of Software Engineering College of Computing Umm Al-Qura University Makkah Saudi Arabia

—Medical Consumer Electronics (MCE) have greatly facilitated people’s lives, allowing patients to obtain personal medical information including medical images. These medical images obtained from MCE need to be processed using segmentation, diagnosis, recognition to obtain useful information. Deep learning plays an important role in medical image segmentation and recognition. However, other personal information about the patient, such as previous cases, is often required to determine the final diagnosis. However, sending personal information and medical images generated through MCE to the clinic directly exposes patient privacy. Therefore, in this paper, we propose a digital medical image privacy protection framework based on information steganography with fuzzy deep learning. The proposed framework transforms patient privacy information into invisible noise and hides it in the deep covert space of the image. This method plays a crucial role in ensuring patient privacy and data security. Furthermore, this method can ensure a high level of image segmentation accuracy that supports the development of personalized medical services. The experimental results show that the proposed framework can embed patient information in medical images without affecting the segmentation task, and effectively protect privacy. Finally, the proposed framework provides new possibilities and impetus for the popularization of MCE. © 2025 IEEE.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

DMSDRec: Dynamic Structure-aware Graph Masked Autoencoder and Spatiotemporal Diffusion for Next-POI Recommendation

引用

IEEE Transactions on Services computing 2025年

作者： Li, Yue Zeng, Jun Tang, Haoran Wen, Junhao Gao, Min Zhou, Wei Chongqing University School of big data and software engineering Chongqing40044 China Hong Kong Polytechnic University Department of Computing Hung Hom Kowloon Hong Kong

The rise of smart devices has accelerated location-based services, creating vast trajectory data for Point-of-Interest (POI) recommendations. However, user omissions or privacy concerns often result in incomplete trajectory data, compromising sequential patterns and spatiotemporal relationships. Existing solutions using graph structures, self-supervised learning (SSL), or spatiotemporal contexts still face two limitations: (1) graph-based and SSL methods produce suboptimal trajectory representations due to respective inherent constraints;(2) noise interference persists when modeling distorted spatiotemporal signals. To mitigate these issues, we propose a Dynamic Structure-aware Graph Masked Autoencoder and Spatiotemporal Diffusion for Next-POI Recommendation (DMSDRec). Specifically, we introduce a dynamic structure-aware improved graph masked autoencoder that adaptively and dynamically distills global transitional information for self-supervised augmentation. It naturally avoids the noise introduced by existing SSL methods' dependency on manual views augmentation. Meanwhile, the masked reconstruction task synergistically enhances trajectory representations by capturing deeper cross-sequence dependencies. Additionally, we propose an effective latent-space spatiotemporal diffusion denoising method. First, we employ graph structures to model spatiotemporal relationships, utilizing higher-order structural information to alleviate the linear spatiotemporal relationship deviations caused by incomplete trajectories. Building on this, we implement diffusion models in the latent space to systematically identify and remove noise from spatiotemporal representations. Through experimental results on two real-world datasets, we demonstrate the superiority of our proposed DMSDRec in terms of recommendation accuracy and robustness. © 2008-2012 IEEE.

关键词： Dynamic Structure-awareness Graph Masked Autoencoder Next POI Recommendation Spatiotemporal Diffusion

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：