检索结果-内蒙古大学图书馆

Drug-Target Interactions Prediction Based on Signed Heterogeneous Graph Neural Networks

Chinese Journal of Electronics 2024年第1期33卷 231-244页

作者： Ming CHEN Yajian JIANG Xiujuan LEI Yi PAN Chunyan JI Wei JIANG College of Information Science and Engineering Hunan Normal University School of Computer Science Shaanxi Normal University Faculty of Computer Science and Control Engineering Shenzhen Institute of Advanced TechnologyChinese Academy of Sciences Computer Science Department BNU-HKBU United International College

Drug-target interactions(DTIs) prediction plays an important role in the process of drug *** computational methods treat it as a binary prediction problem, determining whether there are connections between drugs and targets while ignoring relational types information. Considering the positive or negative effects of DTIs will facilitate the study on comprehensive mechanisms of multiple drugs on a common target, in this work, we model DTIs on signed heterogeneous networks, through categorizing interaction patterns of DTIs and additionally extracting interactions within drug pairs and target protein pairs. We propose signed heterogeneous graph neural networks(SHGNNs), further put forward an end-to-end framework for signed DTIs prediction, called SHGNN-DTI,which not only adapts to signed bipartite networks, but also could naturally incorporate auxiliary information from drug-drug interactions(DDIs) and protein-protein interactions(PPIs). For the framework, we solve the message passing and aggregation problem on signed DTI networks, and consider different training modes on the whole networks consisting of DTIs, DDIs and PPIs. Experiments are conducted on two datasets extracted from Drug Bank and related databases, under different settings of initial inputs, embedding dimensions and training modes. The prediction results show excellent performance in terms of metric indicators, and the feasibility is further verified by the case study with two drugs on breast cancer.

关键词： Drug-target interactions Signed heterogeneous network Link sign prediction Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

BAHBench: A Unified Benchmark for Evaluating Bio-Acoustic Health with Acoustic Foundation Models

引用

IEEE Journal of Biomedical and Health Informatics 2025年第6期PP卷 PP页

作者： Xu, Weixiang Dong, Zhongren Peng, Jing Wang, Runming Zhang, Zixing Hunan Normal University School of Information Science and Engineering Changsha410081 China Hunan University College of Computer Science and Electronic Engineering Changsha410082 China Hunan University Shenzhen Research Institute Shenzhen518000 China

Acoustic foundation models, through self-supervised learning on large amounts of unlabeled speech data, can acquire rich acoustic representations. In recent years, these models have demonstrated substantial potential in audio-based health-related tasks, remarkably enhancing the efficiency and quality of healthcare services and contributing to the advancement of smart healthcare. However, there is currently a lack of systematic research and exploration on the performance of acoustic foundation models in health-related tasks. Furthermore, inconsistencies in evaluation methods and experimental setups hinder fair comparisons between different methods, severely impeding progress in this field. To address these challenges, we establish a unified Benchmark for evaluating Bio-Acoustic health via acoustic foundation models, namely BAHBench. BAHBench encompasses 6 distinct health-related tasks and evaluates 12 acoustic foundation models within a unified evaluation framework and parameter settings, enabling fair comparisons across different models. Our objective is to explore the effectiveness of current acoustic foundation models in health-related tasks. Thus, we discuss the impact of model size and data diversity on performance, and investigate feature selection and efficient fine-tuning strategy. Experimental results show that different health-related tasks benefit from features from different layers of the foundation model, while LoRA fine-tuning further enhances the model's performance on downstream tasks. Our goal is to provide clear and comprehensive guidance for future researchers. The code related to this study will be available to the research community to promote transparency and reproducibility. © 2013 IEEE.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Fake Face Detection Based on Fusion of Spatial Texture and High-Frequency Noise

引用

Chinese Journal of Electronics 2025年第1期34卷 212-221页

作者： Dengyong Zhang Feifan Qi Jiahao Chen Jiaxin Chen Rongrong Gong Yuehong Tian Lebing Zhang Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation School of Computer and Communication Engineering Changsha University of Science and Technology School of Computer and Communication Engineering Changsha University of Science and Technology Changsha Social Work College Changkuangao Beijing Technology Co. Ltd. School of Computer and Artificial Intelligence Huaihua University

The rapid development of the Internet has led to the widespread dissemination of manipulated facial images, significantly impacting people's daily lives. With the continuous advancement of Deepfake technology, the generated counterfeit facial images have become increasingly challenging to distinguish. There is an urgent need for a more robust and convincing detection method. Current detection methods mainly operate in the spatial domain and transform the spatial domain into other domains for analysis. With the emergence of transformers, some researchers have also combined traditional convolutional networks with transformers for detection. This paper explores the artifacts left by Deepfakes in various domains and, based on this exploration, proposes a detection method that utilizes the steganalysis rich model to extract high-frequency noise to complement spatial features. We have designed two main modules to fully leverage the interaction between these two aspects based on traditional convolutional neural networks. The first is the multi-scale mixed feature attention module, which introduces artifacts from high-frequency noise into spatial textures, thereby enhancing the model's learning of spatial texture features. The second is the multi-scale channel attention module, which reduces the impact of background noise by weighting the features. Our proposed method was experimentally evaluated on mainstream datasets, and a significant amount of experimental results demonstrate the effectiveness of our approach in detecting Deepfake forged faces, outperforming the majority of existing methods.

关键词： Deepfakes Adaptation models Attention mechanisms Noise Transforms Transformers Feature extraction Internet Convolutional neural networks Faces

来源：评论

学校读者我要写书评

暂无评论

Towards Source-Free Domain Adaptive Semantic Segmentation Via Importance-Aware and Prototype-Contrast Learning

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-13页

作者： Cao, Yihong Zhang, Hui Lu, Xiao Xiao, Zheng Yang, Kailun Wang, Yaonan College of Computer Science and Electronic Engineering Hunan University Changsha China National Engineering Research Center of Robot Vision Perception and Control Technology School of Robotics Hunan University Changsha China College of Engineering and Design Hunan Normal University Changsha China

Domain adaptive semantic segmentation enables robust pixel- wise understanding in real-world driving scenes. Source-free domain adaptation, as a more practical technique, addresses the concerns of data privacy and storage limitations in typical unsupervised domain adaptation methods, making it especially relevant in the context of intelligent vehicles. It utilizes a well-trained source model and unlabeled target data to achieve adaptation in the target domain. However, in the absence of source data and target labels, current solutions cannot sufficiently reduce the impact of domain shift and fully leverage the information from the target data. In this paper, we propose an end-to-end source-free domain adaptation semantic segmentation method via Importance-Aware and Prototype-Contrast (IAPC) learning. The proposed IAPC framework effectively extracts domain-invariant knowledge from the well-trained source model and learns domain-specific knowledge from the unlabeled target domain. Specifically, considering the problem of domain shift in the prediction of the target domain by the source model, we put forward an importance-aware mechanism for the biased target prediction probability distribution to extract domain-invariant knowledge from the source model. We further introduce a prototype-contrast strategy, which includes a prototype-symmetric cross-entropy loss and a prototype-enhanced cross-entropy loss, to learn target intra-domain knowledge without relying on labels. A comprehensive variety of experiments on two domain adaptive semantic segmentation benchmarks demonstrates that the proposed end-to-end IAPC solution outperforms existing state-of-the-art methods. The source code is publicly available at https://***/yihong-97/Source-free-IAPC. IEEE

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

An improving integration-enhanced ZNN for solving time-varying polytope distance problems with inequality constraint

引用

Neural Computing and Applications 2024年第29期36卷 18237-18250页

作者： Li, Hao Zhang, Zhen Liao, Bolin Hua, Cheng College of Computer Science and Engineering Jishou University Hunan Jishou416000 China School of Communication and Electronic Engineering Jishou University Hunan Jishou416000 China School of Electronic Information and Electrical Engineering Changsha university Hunan Changsha410073 China

Time-varying polytope distance (TVPD) problems are prevalent in scientific and engineering applications and can be transformed into time-varying quadratic programming (TVQP) problems with both equality and inequality constraints. Concurrently, the noise interferences during the solution process are non-negligible and challenging to eliminate. Although zeroing neural networks (ZNNs) perform well in solving various types of time-varying problems, they still fall short in the suppression of unbounded noises, such as linear noise. To address this limitation, this paper proposes an improving integration-enhanced ZNN (IIEZNN) model for accurately solving TVPD problems under noise environments. Compared with the existing ZNN models, the IIEZNN model has stronger inherent robustness. The stability and robustness of the IIEZNN model are guaranteed by rigorous theoretical analysis. Firstly, the effectiveness of the IIEZNN model is verified via two TVQP examples. Then, the IIEZNN model is generalized to TVPD problem solving and has excellent performance. Specifically, in solving the TVPD under linear noises, the residual error of the IIEZNN model converges to the order of 10-5, which is much lower than that of the existing noise-tolerant ZNN model with an order of 10-1. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Quadratic programming

来源：评论

学校读者我要写书评

暂无评论

Towards Privacy Protection: A Non-Invasive Human Detection Method via Edge-Cloud Collaboration 10

Towards Privacy Protection: A Non-Invasive Human Detection M...

引用

10th IEEE Conference on Big Data Security on Cloud, BigDataSecurity 2024

作者： Yin, Luxiu Huang, Qihui Wang, Xuanya Liang, Wei Wang, Chun Hunan University of Science and Technology College of Computer Science and Engineering China Hunan Normal University College of Information Science and Engineering China

ISBN: (纸本)9798350389524

Traditional human detection technologies, such as those based on computer vision or wearable devices, often require the storage and analysis of vast amounts of personal data in the cloud, which increases the risk of data breaches. Therefore, the rapid processing of large-scale detection data while protecting user privacy has become a critical issue that needs urgent resolution. The approaches based on wireless sensing will achieve privacy protection, non-contact perception, massively scalable, and long-range detection. In this paper, we propose an innovative solution that combines data collection from terminal devices, real-time processing by edge computing, and deep analysis capabilities of cloud computing, to realize efficient and privacy-friendly human detection. First, we propose a non-invasive human detection method based on edge-cloud collaboration, integrating wireless sensing, data transmission, and processing within edge nodes, and utilizing drones equipped with edge nodes to process a large amount of wireless sensing data in real-time. Secondly, we designed a human presence prediction module based on a lightweight convolutional neural network to analyze signal data and achieve non-invasive human detection. Finally, with commodity software defined radio devices, we implement a non-invasive human detection prototype that can accurately differentiate multiple states of targets while protecting their privacy. Experimental results indicate that the proposed solution effectively reduces the risk of privacy leakage while maintaining a high detection accuracy. This provides strong support for the application of non-invasive personnel detection technology in the aspect of privacy protection. © 2024 IEEE.

关键词： Digital storage

来源：评论

学校读者我要写书评

暂无评论

Deep learning for iris recognition: a review

引用

Neural Computing and Applications 2025年第17期37卷 11125-11173页

作者： Yin, Yimin He, Siliang Zhang, Renye Chang, Hongli Zhang, Jinghua School of Mathematics and Statistics Hunan First Normal University Hunan Changsha410205 China School of Computer Science Hunan First Normal University Changsha410205 China School of Information Science and Engineering Southeast University Nanjing210096 China College of Intelligence Science and Technology National University of Defense Technology Hunan Changsha410073 China

Iris recognition is a secure biometric technology known for its stability and privacy. With no two irises being identical and little change throughout a person’s lifetime, iris recognition is considered more reliable and less susceptible to external factors than other biometric recognition methods. Unlike traditional machine learning-based iris recognition methods, deep learning technology does not rely on feature engineering and boasts excellent performance. This paper collects 131 relevant papers to summarize the development of iris recognition based on deep learning. We introduce the background of iris recognition and the motivation and contribution of this survey. Then, we present the common datasets widely used in iris recognition. After that, we summarize the key tasks involved in the process of iris recognition based on deep learning technology, including identification, segmentation, presentation attack detection, and localization. Finally, we discuss the challenges and potential development of iris recognition. This review provides a comprehensive sight of the research of iris recognition based on deep learning. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Graph Neural Networks for Drug Response Prediction 9

Graph Neural Networks for Drug Response Prediction

引用

9th International Conference on Systems and Informatics, ICSAI 2023

作者： Zhu, Jiehuang Liu, Xianfeng Zhang, Jin Wu, Sheng Hunan Normal University College of Information and Computer Engineering Changsha China Changsha University of Science and Technology School of Computer and Communication Engineering Changsha China

ISBN: (纸本)9798350383706

Drug response prediction plays a crucial role in precision medicine, such as cancer analysis and treatment. Due to the uncertainty of drug efficacy and the heterogeneity of cancer, predicting drug response in vitro is expected to assist in guiding anticancer drug design and understanding cancer biology. Based on the Genomics of Drug Sensitivity in Cancer (GDSC) and the Cancer Cell Line Encyclopedia (CCLE), we propose a deep learning method for predicting drug sensitivity. Specifically, we first construct graphs of drugs and cell lines in the benchmark dataset. And we design a drug feature extraction module based on Graph Transformer and a cell line feature extraction module based on Graph Attention Networks (GAT) to capture the embeddings of drugs and cell lines. Finally, we feed the drug embeddings and cell line embeddings into the module for prediction. Experiment results show that our method outperforms several advanced models in drug response prediction, demonstrating the good performance and enormous application potential in precision medicine. In addition, the results of ablation experiments demonstrate the effectiveness of the modules in our model. © 2023 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

A Novel Secure Scan Design Based on Delayed Physical Unclonable Function

引用

computers, Materials & Continua 2023年第3期74卷 6605-6622页

作者： Weizheng Wang Xingxing Gong Xiangqi Wang Gwang-jun Kim Fayez Alqahtani Amr Tolba School of Computer and Communication Engineering Changsha University of Science and TechnologyChangsha410114China College of Information Science and Engineering Hunan Women’s UniversityChangsha410138China College of Mathematics and Computational Science Hunan First Normal UniversityChangsha410138China Department of Computer Engineering Chonnam National UniversityGwangju61186Korea Software Engineering Department College of Computer and Information SciencesKing Saud UniversityRiyadh12372Saudi Arabia Computer Science Department Community CollegeKing Saud UniversityRiyadh11437Saudi Arabia

The advanced integrated circuits have been widely used in various situations including the Internet of Things,wireless communication,*** its manufacturing process exists unreliability,so cryptographic chips must be rigorously *** to scan testing provides high test coverage,it is applied to the testing of cryptographic integrated ***,while providing good controllability and observability,it also provides attackers with a backdoor to steal *** the text,a novel protection scheme is put forward to resist scan-based attacks,in which we first use the responses generated by a strong physical unclonable function circuit to solidify fuseantifuse structures in a non-linear shift register(NLSR),then determine the scan input code according to the configuration of the fuse-antifuse structures and the styles of connection between the NLSR cells and the scan *** the key is right,the chip can be tested normally;otherwise,the data in the scan chain cannot be propagated normally,it is also impossible for illegal users to derive the desired scan *** proposed technique not only enhances the security of cryptographic chips,but also incurs acceptable overhead.

关键词： Cryptographic chips scan testing scan-based attacks hardware security PUF

来源：评论

学校读者我要写书评

暂无评论

RDP: Ranked Differential Privacy for Facial Feature Protection in Multiscale Sparsified Subspaces

引用

IEEE Internet of Things Journal 2025年第11期12卷 17333-17348页

作者： Ou, Lu Liao, Shaolin Gao, Shihui Huang, Guandong Qin, Zheng Hunan University School of Journalism and Communication Hunan Changsha410082 China Sun Yat-sen University School of Electronics and Information Technology Guangzhou510275 China Illinois Institute of Technology Department of Electrical and Computer Engineering ChicagoIL60616 United States Purdue University Elmore School of Electrical and Computer Engineering West LafayetteIN47907 United States Hunan University College of Computer Science and Electronic Engineering Hunan Changsha410082 China State Grid Hunan Materials Company Department of Deployment Management Hunan Changsha410118 China

With the widespread sharing of personal face images on the Internet, systems based on face recognition encounter the real threat of being breached by potential adversaries who are able to access individuals’ face images and use them to intrude the systems. In this article, we propose a novel privacy protection method in the sparsified multiscale feature subspaces to protect sensitive facial features, taking care of the influence or weight-ranked subspace coefficients on the privacy budget, named "ranked differential privacy (RDP)." After the multiscale subspaces’ decomposition, the lightweight Laplacian noise is added to the dimension-reduced sparsified subspaces’ coefficients according to the geometric superposition method. Then, we rigorously prove that the RDP satisfies Ε0-differential privacy. After that, the nonlinear Lagrange multiplier method (LM) is formulated for the constraint optimization problem of maximizing the utility of protected face images of high-visualization quality with sanitizing noise, under a given privacy budget Ε0. Then, two methods are proposed to solve the nonlinear Lagrangian multiplier method (LM) problem and obtain the optimal noise scale parameters: 1) the analytical normalization approximation (NA) method with identical average noise scale parameter for real-time online applications and 2) the LM optimization method via gradient descent (LMGD) to obtain the nonlinear solution through iterative updating for more accurate offline applications. Experimental results on two real-world datasets show that our proposed RDP outperforms other state-of-the-art methods: at a privacy budget of Ε0 = 0.2 , the peak signal-to-noise ratio (PSNR) of the optimized RDP is about ~10 dB higher than (10 times as high as) the highest PSNR of all state-of-the-art methods compared. © 2014 IEEE.

关键词： Differential privacy

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：