检索结果-内蒙古大学图书馆

IEEE Transactions on technology and Society 2024年第2期5卷 217-230页

作者： Pangelinan, Gabriella Bhatta, Aman Wu, Haiyu King, Michael C. Bowyer, Kevin W. Florida Institute of Technology Department of Electrical Engineering and Computer Science MelbourneFL32901 United States University of Notre Dame Department of Computer Science and Engineering Notre DameIN46556 United States

Concerns related to the proliferation of automated face recognition technology with the intent of solving or preventing crimes continue to mount. The technology being implicated in wrongful arrest following 1-to-many face identification has heightened scrutiny in the media and spawned lawsuits. We analyze the accuracy of 1-to-many face identification across African-American and Caucasian demographic groups and in the presence of blur and reduced resolution in the probe image, as might occur in "surveillance camera quality"images. Because of the high accuracy of modern face recognition algorithms with "government ID quality"images and the size of available datasets, we use the following indirect metrics to assess the relative propensity for false positive identifications: (1) the traditional d-prime statistic between mated and non-mated score distributions, (2) absolute score difference between thresholds in the high-similarity tail of the non-mated distribution and the low-similarity tail of the mated distribution, and (3) distribution of (mated - non-mated rank-one scores) across the set of probe images. We find that demographic variation patterns in 1-to-many accuracy largely, but not perfectly, follow that observed in 1-to-1 accuracy. We show that, different from the case with 1-to-1 matching accuracy, demographic comparison of 1-to-many accuracy can be affected by different numbers of identities and images across demographics. We also show that increased blur or reduced resolution of the face in the probe image can significantly increase the false positive identification rate. This point is important because the reputation of modern face recognition algorithms for high accuracy in 1-to-many face identification comes from statistics based on analyzing "government ID quality"rather than "surveillance camera quality"images. © 2024 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Multi-factor Authentication and Data Integrity for WBAN Using Hash-Based Techniques 5th

Multi-factor Authentication and Data Integrity for WBAN Usin...

引用

5th Doctoral Symposium on Computational Intelligence, DoSCI 2024

作者： Pabitha, B. Vani, V. Sanshi, Shridhar Karthik, N. Department of Computer Science and Engineering National Institute of Technology Puducherry Karaikal India Department of Computer Science and Engineering National Institute of Technology Karnataka Mangalore India

ISBN: (纸本)9789819767250

In recent days, a wireless body area network (WBAN) has been developed as part of the Internet of Things (IoT) with sensors and actuators in three different modes, building its network, i.e., in-body sensors, wearable sensors, and on-body sensors. The doctor’s access the data recorded and monitored by the sensor embedded in the patient to treat critical situations immediately. Maintaining data integrity and guarding against threats is necessary to secure sensitive patient information. Several people have proposed schemes for authenticating data access through formal and informal verification. In this research work, we carry out multi-factor authentication extensively using zero-knowledge proofs. The anomaly detection of the sensors is detected using machine learning algorithms, which help tune the sensors to their correct working conditions. The work aims to concentrate on sensor working conditions promptly and to handle attacks like masquerade, forgery, and key escrow attacks. To assess whether performance metrics are superior in computing cost, storage overhead, and communication overhead, utilize the BAN logic tool. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Wearable sensors

来源：评论

学校读者我要写书评

暂无评论

emapDiffP: A novel learning algorithm for convolutional neural network optimization

引用

Neural Computing and Applications 2024年第20期36卷 11987-12010页

作者： Bhakta, Shubhankar Nandi, Utpal Changdar, Chiranjit Ghosal, Sudipta Kr Pal, Rajat Kumar Department of Computer Science Vidyasagar University West Bengal Midnapore721102 India Department of Computer Science Belda College West Bengal Belda721424 India Department of Computer Science and Technology Behala Goverment Polytechnic West Bengal Kolkata720060 India Department of Computer Science and Engineering University of Calcutta West Bengal Kolkata700106 India

Deep neural networks (DNNs) having multiple hidden layers are very efficient to learn large volume datasets and applied in a wide range of applications. The DNNs are trained on these datasets using learning algorithms to learn the relationships among different variables. The base method that makes DNNs successful is stochastic gradient descent (SGD). The gradient reveals the way that a function’s steepest rate of alteration is occurring. No matter how the gradient behaves, the key issue with basic SGD is that all parameters must adjust in equal-sized increments. Consequently, creating adaptable step sizes for every parameter is an effective method of deep model optimization. Gradient-based adaptive techniques utilize local changes in gradients or the square roots of exponential moving averages of squared previous gradients. However, current optimizers continue to struggle with effectively utilizing optimization curved knowledge. The novel emapDiffP optimizer suggested in this study utilizes the prior two parameters to generate a non-periodic and non-negative function, and the upgrade parameter makes use of a partly adaptive value to account for learning rate adjustability. Thus, the optimization steps become smoother with a more accurate step size for the immediate past parameter, a partial adapting value, and the largest two momentum values as the denominator of parameter updating. The rigorous tests on benchmark datasets show that the presented emapDiffP performs significantly better than its counterparts. In terms of classification accuracy, the emapDiffP algorithm gives the best classification accuracy on CIFAR10, MNIST, and Mini-ImageNet datasets for all examined networks and on the CIFAR100 dataset for most of the networks examined. It offers the best classification accuracy on the ImageNet dataset with the ResNet18 model. For image classification tasks on various datasets, the suggested emapDiffP technique offers outstanding training speed. With MNIST, CIFAR1

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Leveraging XGBoost and Clinical Attributes for Heart Disease Prediction 5

Leveraging XGBoost and Clinical Attributes for Heart Disease...

引用

5th International Conference on Sustainable Communication Networks and Application, ICSCNA 2024

作者： Susmitha, Kota Bhargavi, Maridu Sri, Achyuta Mohitha Sai Reddy, Bogala Devi Prasaad Satyanarayana, Paladugu Siva Rao, K V Ranga Vignan's Foundation for Science Technology and Research Department of Computer Science and Engineering Guntur India Department of Computer Science and Engineering-Data Science Hyderabad India

ISBN: (纸本)9798331530013

Despite the various advances in medical technology, heart disease continues to rank among the leading causes of mortality in the world, killing millions each year. There is hope that the risks involved with heart disease can be reversed in time and the outcomes of patients improved with earlier prediction and diagnosis. However, the use of conventional methods of diagnostic techniques tends to be very resource-intensive and inaccessible to other more disadvantaged areas. The present work defines a model for machine learning in predicting heart disease and illustrates a more efficient and accessible solution. There is a risk for heart disease that is determined using a set of various clinical features. The algorithm used in the prediction of heart disease is XG Boost that are applied to the dataset. The method provides a reliable tool for the early detection of heart disease, thus helping healthcare providers to improve preventive care and treatment strategies. The results of this study will be helpful in advancing research to identify other avenues for prediction of heart disease. © 2024 IEEE.

关键词： early diagnosis Heart disease UCI (University California Irvine) Heart Disease dataset XG Boost

来源：评论

学校读者我要写书评

暂无评论

Neural Network-Driven Base Station Selection in Next-Generation Cellular Networks 17

Neural Network-Driven Base Station Selection in Next-Generat...

引用

17th International Conference on COMmunication Systems and NETworkS, COMSNETS 2025

作者： Harrithha, B. Kotagi, Vijeth J. Thakur, Rahul Indian Institute of Technology Dharwad Department of Computer Science and Engineering Karnataka Dharwad India Indian Institute of Technology Roorkee Department of Computer Science and Engineering Uttarakhand Roorkee India

ISBN: (纸本)9798331531195

With the exponential increase in data consumption from activities like streaming, gaming, and IoT applications, existing network infrastructure is under significant strain. To address the growing demand for higher data rates and better coverage, increasing base station density and exploring data offloading in unlicensed spectrum have become essential. While higher base station density reduces the distance between devices and network nodes, enhancing signal strength and data throughput, it also complicates the process of base station selection for users, particularly for those at the cell edge. Thus, ever-increasing demand for high-speed, low-latency communication services in 5G and 5G advanced New Radio (NR) unlicensed (NR-U) networks necessitates intelligent and dynamic approaches for user admission control and resource *** paper presents a novel solution employing neural network technology to address these critical challenges. We first propose an efficient algorithm for admission control and resource allocation. Upon generation of training data using the above algorithm, we propose a neural network-based framework for cellular networks in unlicensed spectrum such as LTE/LTE-U and 5G/NR-U networks. © 2025 IEEE.

关键词： Base stations

来源：评论

学校读者我要写书评

暂无评论

An Adaptive IDS Scheduling Framework for Host Based Mimicry Attack Detection in IoT Networks 18

An Adaptive IDS Scheduling Framework for Host Based Mimicry ...

引用

18th IEEE International Conference on Advanced Networks and Telecommunications Systems, ANTS 2024

作者： Sanyal, Surja Khatua, Manas Chattopadhyay, Pratik Indian Institute of Technology Guwahati Department of Computer Science and Engineering Guwahati India Varanasi Department of Computer Science and Engineering Varanasi India

ISBN: (纸本)9798350391725

With the expanding applicability of the Internet of Things (IoT), novel IoT network security challenges also appear more frequently. Host-based Mimicry Attacks (HMA) are one of them that is difficult to detect by traditional Intrusion Detection Systems (IDS) since they mimic legitimate network and host behaviors to evade detection. Additionally, IDSs consume host device resources when their implementations, partially or fully, are executed in the device, and thus it is needed to be scheduled properly so as to save device resources when possible. Current IDS solutions capable of detecting HMAs mostly use randombased scheduling or do not use scheduling at all and so are always executed. Random-based scheduling solutions have low HMA detection accuracy, while solutions without scheduling do not save device resources. This work proposes an adaptive IDS scheduling framework that suspends IDS execution based on device trust and resumes IDS execution when the device is under attack. Extensive experimentation using the Cooja simulator and Python script shows that the proposed framework achieves a device behavior classification accuracy of at least 97.15% when under attack while saving at least 81.98% of CPU energy when not under attack for a network of 40 devices, as compared to those when executing the IDS without any scheduling scheme. © 2024 IEEE.

关键词： Network intrusion

来源：评论

学校读者我要写书评

暂无评论

An empirical evaluation of teaching–learning-based optimization, genetic algorithm and particle swarm optimization

引用

International Journal of computers and Applications 2023年第1期45卷 36-50页

作者： Shukla, Alok Kumar Pippal, Sanjeev Kumar Chauhan, Sansar Singh Department of Computer Science and Engineering G.L. Bajaj Institute of Technology & Management Greater Noida India Department of Computer Science & Engineering Galgotias University Greater Noida India

Metaheuristic algorithms have probable to solve global optimization problems in various fields of engineering and industry. To find a global solution by exploring irregular or non-linear surfaces, classical optimization techniques are not able. To overcome the limitation of the classical approach, in the recent study, a large number of metaheuristic methods have been investigated to improve the solution quality concerning convergence and accuracy on the complex problems. Nowadays, a popular metaheuristic algorithm is introduced called teaching–learning-based optimization (TLBO). It is recently being used as an innovative, and robust method to solve the global optimization problem, inspired by the teaching–learning phenomenon. On the other hand, particle swarm optimization (PSO) algorithm is one of the most utilized algorithms in the current scenario, which has indicated acceptable efficiency. Genetic algorithm (GA) is a valuable component of metaheuristic which has been applied in various research applications. In this article, the performance of TLBO algorithm is estimated on 25 numerical test suites against other metaheuristic algorithms such as PSO and GA. The results of our experimental study show that TLBO outperforms the PSO and GA algorithm regarding convergence solution. © 2019 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning based Adaptive Healthcare Decision Support Systems using Time Series Forecasting 5

Reinforcement Learning based Adaptive Healthcare Decision Su...

引用

5th International Conference on Data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Shirley, C.P. Berin Jeba Jingle, I. Abisha, M.B. Venkadesan, R. Absin, S.J. Karunya Institute of Technology and Sciences Department of Computer Science Engineering Coimbatore India Karunya Institute of Technology and Sciences Department of Computer Science and Engineering Coimbatore India

ISBN: (数字)9798350389609

ISBN: (纸本)9798350389609

Healthcare decision-making is a complex and crucial process that must take into account changing patient situations and medical trends. It is frequently difficult for conventional decision support systems to adjust to these shifting circumstances. In this work, we present a novel method for creating adaptive healthcare decision support systems by combining time series forecasting and reinforcement learning (RL). The main goal is to create a system that can use reinforcement learning (RL) to develop optimal policies for making decisions by utilizing time series data of patients from the past. Our system uses RL in conjunction with time series forecasting methods like LSTM (Long Short-Term Memory) networks and ARIMA (Auto Regressive Integrated Moving Average) to estimate future patient outcomes and healthcare resource requirements. There are several different components in the system architecture. The time series forecasting models are trained using historical healthcare data, which includes patient demographics, medical histories, and treatment outcomes. Forecasts for important healthcare parameters like patient recovery rates, disease progression, and resource consumption are produced by these models. According to the anticipated results, RL algorithms are used to learn decision policies. While interacting with a simulated environment that represents patient care scenarios, the RL agent aims to maximize long-term rewards, which may include reduced expenses, better patient outcomes, or more efficient use of available resources. Using actual healthcare datasets, the adaptability and efficacy of suggested approach will be assessed. The efficacy of RL-based decision policies, accuracy of projected results, and the system's adaptability to shifting healthcare dynamics are all examples of performance measures. © 2024 IEEE.

关键词： Decision support systems Adaptive systems Accuracy Time series analysis Systems architecture Medical services Reinforcement learning Predictive models Forecasting Long short term memory

来源：评论

学校读者我要写书评

暂无评论

AttrLeaks on the Edge: Exploiting Information Leakage from Privacy-Preserving Co-inference

引用

Chinese Journal of Electronics 2023年第1期32卷 1-12页

作者： WANG Zhibo LIU Kaixin HU Jiahui REN Ju GUO Hengchang YUAN Wei School of Cyber Science and Engineering Wuhan University School of Cyber Science and Technology Zhejiang University Department of Computer Science and Technology Tsinghua University

Collaborative inference(co-inference) accelerates deep neural network inference via extracting representations at the device and making predictions at the edge server, which however might disclose the sensitive information about private attributes of users(e.g.,race). Although many privacy-preserving mechanisms on co-inference have been proposed to eliminate privacy concerns, privacy leakage of sensitive attributes might still happen during inference. In this paper, we explore privacy leakage against the privacy-preserving co-inference by decoding the uploaded representations into a vulnerable form. We propose a novel attack framework named AttrL eaks, which consists of the shadow model of feature extractor(FE), the susceptibility reconstruction decoder,and the private attribute classifier. Based on our observation that values in inner layers of FE(internal representation) are more sensitive to attack, the shadow model is proposed to simulate the FE of the victim in the blackbox scenario and generates the internal ***, the susceptibility reconstruction decoder is designed to transform the uploaded representations of the victim into the vulnerable form, which enables the malicious classifier to easily predict the private attributes. Extensive experimental results demonstrate that AttrLeaks outperforms the state of the art in terms of attack success rate.

关键词： Deep learning Privacy Collaboration Transforms Feature extraction Prediction algorithms Iron

来源：评论

学校读者我要写书评

暂无评论

RAPID-AV: Realtime Adaptive and Intelligent Detection For AV's Against DDoS Attacks 18

RAPID-AV: Realtime Adaptive and Intelligent Detection For AV...

引用

18th IEEE International Conference on Advanced Networks and Telecommunications Systems, ANTS 2024

作者： Kotha, Aditya Venkanna, U. International Institute of Information Technology Department of Computer Science and Engineering Naya Raipur India National Institute of Technology Department of Computer Science and Engineering Warangal India

ISBN: (纸本)9798350391725

Autonomous vehicles (AVs) increasingly rely on vehicle-to-everything (V2X) networks for communication. However, due to the devices' heterogeneity, they are more susceptible to attacks like distributed denial of service (DDoS), resulting in packet loss and transmission delays. Existing solutions to these issues often create bottlenecks and lack post-attack recovery mechanisms. To solve these problems, this paper proposes RAPID-AV, an ML-based approach that uses programmable P4 switches for real-time detection and mitigation of DDoS attacks in AV networks. The proposed system uses a threefold strategy: collect traffic statistics from AV, use P4-based programmable packet processing for anomaly detection, and use majority voting to mitigate attacks. In our experiments, we found that the decision tree algorithm achieves an accuracy of 94.33% in distinguishing traffic from DDoS traffic. In addition, RAPID-AV demonstrated a nearly 10% reduction in CPU load during attacks and maintained consistent communication between AV and RSU. This approach can help improve AV network security while addressing the limitations of existing solutions. © 2024 IEEE.

关键词： Network security

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：