检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Khot, Ayush Wang, Xiwei Roy, Avik Kindratenko, Volodymyr Neubauer, Mark S. Department of Physics University of Illinois Urbana-Champaign UrbanaIL61801 United States The Grainger College of Engineering Department of Electrical and Computer Engineering University of Illinois Urbana-Champaign UrbanaIL61801 United States Center for Artificial Intelligence Innovation National Center for Supercomputing Applications University of Illinois Urbana-Champaign UrbanaIL61801 United States The Grainger College of Engineering Siebel School of Computing and Data Science University of Illinois Urbana-Champaign UrbanaIL61801 United States

Current methods commonly used for uncertainty quantification (UQ) in deep learning (DL) models utilize Bayesian methods which are computationally expensive and time-consuming. In this paper, we provide a detailed study of UQ based on evidential deep learning (EDL) for deep neural network models designed to identify jets in high energy proton-proton collisions at the Large Hadron Collider and explore its utility in anomaly detection. EDL is a DL approach that treats learning as an evidence acquisition process designed to provide confidence (or epistemic uncertainty) about test data. Using publicly available datasets for jet classification benchmarking, we explore hyperparameter optimizations for EDL applied to the challenge of UQ for jet identification. We also investigate how the uncertainty is distributed for each jet class, how this method can be implemented for the detection of anomalies, how the uncertainty compares with Bayesian ensemble methods, and how the uncertainty maps onto latent spaces for the models. Our studies uncover some pitfalls of EDL applied to anomaly detection and a more effective way to quantify uncertainty from EDL as compared with the foundational EDL setup. These studies illustrate a methodological approach to interpreting EDL in jet classification models, providing new insights on how EDL quantifies uncertainty and detects out-of-distribution data which may lead to improved EDL methods for DL models applied to classification tasks. © 2025, CC BY.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Analyzing AI regulation through literature and current trends

引用

Journal of Open Innovation: Technology, Market, and Complexity 2025年第1期11卷

作者： Shetty, Dasharathraj K. Vijaya Arjunan, R. Cenitta, D. Makkithaya, Krishnamoorthi Hegde, Nikhil Venkatraman Bhatta B, Shreepathy Ranga Salu, Staissy Aishwarya, T.R. Bhat, Pranav Pullela, Phani Kumar Department of Data Science and Computer Applications Manipal Institute of Technology Manipal Academy of Higher Education (MAHE) Karnataka Manipal 576104 India Department of Computer Science and Engineering Manipal Institute of Technology Manipal Academy of Higher Education (MAHE) Karnataka Manipal 576104 India Department of Humanities and Management Manipal Institute of Technology Manipal Academy of Higher Education (MAHE) Karnataka Manipal 576104 India Department of Biomedical Engineering Manipal Institute of Technology Manipal Academy of Higher Education (MAHE) Karnataka Manipal 576104 India Department of Hospital Administration Kasturba Medical College Manipal Academy of Higher Education (MAHE) Karnataka Manipal 576104 India School of Computer Science and Engineering RV University RV Vidyanikethan Post 8th Mile Mysore Rd Mailasandra Karnataka Bengaluru 560059 India

Artificial intelligence (AI) technology has been rapidly evolving, necessitating that discussions on the need and design of regulatory frameworks be taken seriously. This paper reviews literature regarding AI regulations on a theoretical versus a practical level. It examines the different models of regulation – some of which are risk-based regulation and others are complete prohibition – to gauge the literature's predictions on the scope, precise features, and direction of AI regulation. With particular reference to the EU AI Act as the primary case study, the paper analyzes the consequences of rules on innovation and the international standards’ impact on AI regulatory measures. The results show that almost all scholarly assumptions are accurate, showing how the issues are integrated at the practical level and where is it still challenging to comply and enforce. This paper advocates for continuous improvement in AI regulatory frameworks and more international interactions to ensure efficient governance for AI technologies. It contributes to ongoing debates about creating future-proof adaptable robust AI regulations which can negotiate the complexity between technological development and social protection. © 2025 The Authors

关键词： AI Governance mechanisms AI regulation AI Regulatory frameworks Artificial Intelligence policies Comparative study AI regulation EU AI Act Future trends AI regulation

来源：评论

学校读者我要写书评

暂无评论

Polygene-based evolutionary algorithms with frequent pattern mining

引用

Frontiers of computer science 2018年第5期12卷 950-965页

作者： Shuaiqiang WANG Yilong YIN Research Center of Big Data Applications Qilu University of Technology Jinan 250100 China Department of Computer Science and Information Systems University of Jyvaskyla Jyvaskyla 40014 Finland School of Computer Science and Technology Shandong University Jinan 250101 China

In this paper, we introduce polygene-based evolution, a novel framework for evolutionary algorithms （EAs） that features distinctive operations in the evolutionary process. In traditional EAs, the primitive evolution unit is a gene, wherein genes are independent components during evolution. In polygene-based evolutionary algorithms （PGEAs）, the evolution unit is a polygene, i.e., a set of co-regulated genes. Discovering and maintaining quality polygenes can play an effective role in evolving quality individuals. Polygenes generalize genes, and PGEAs generalize EAs. Implementing the PGEA framework involves three phases：（Ⅰ） polygene discovery, （Ⅱ） polygene planting, and （Ⅲ） polygene-compatible evolution. For Phase I, we adopt an associative classificationbased approach to discover quality polygenes. For Phase Ⅱ, we perform probabilistic planting to maintain the diversity of individuals. For Phase Ⅲ, we incorporate polygenecompatible crossover and mutation in producing the next generation of individuals. Extensive experiments on function optimization benchmarks in comparison with the conventional and state-of-the-art EAs demonstrate the potential of the approach in terms of accuracy and efficiency improvement.

关键词： polygenes evolutionary algorithms function optimization associative classification data mining

来源：评论

学校读者我要写书评

暂无评论

An efficient optimized deep learning model for the risk prediction of arterial stiffness in diabetes patients

引用

Biomedical Signal Processing and Control 2025年 110卷

作者： A. Mohana Priya P. Rajesh Kanna S. Vanithamani G. Sathish Kumar Department of Computer Science and Engineering Sathyabama Institute of Science and Technology Chennai Tamil Nadu India Department of Computer Science and Engineering Bannari Amman Institute of Technology Erode Tamil Nadu India Department of Master of Computer Applications M.Kumarasamy College of Engineering Karur Tamil Nadu India Department of Artificial Intelligence and Data Science Sri Eshwar College of Engineering Coimbatore Tamil Nadu India

Arterial Stiffness (AS) is a potential biomarker for assessing cardiovascular disease (CVD) risk in patients with diabetes. It is commonly evaluated using Pulse Wave Velocity (PWV) where increased arterial stiffness is a major contributor to CVD-related events. However, the relationship between cardiovascular risk and diabetes remains unclear and machine learning based clinical prediction models for AS detection are still underdeveloped. Monitoring AS can provide early insights into abnormalities in diabetic patients enabling timely intervention. This study proposes an optimized deep learning model to assess AS risk in diabetic patients to improve CVD risk prediction. The model utilizes brachial-ankle pulse wave velocity (baPWV) and Fasting Blood Glucose (FBG) as diagnostic parameters. First, critical features from patient medical records are extracted using the Least Absolute Shrinkage and Selection Operator (LASSO) method followed by Principal Component Analysis (PCA) to remove outliers using Mahalanobis distance and reconstruction error. Next, a feature selection process is applied using the Gaussian LASSO Cat Optimization Algorithm (GLCOA). Finally, the Deep Kronecker Neural Network (DKNN) optimized with the Frog Leap Algorithm (FLA) referred to as DKNN-FL. In addition to this, SMOTE augmentation and cross validation are applied to avoid overfitting. The statistical test ANOVA shows the proposed work gives a better performance on a discovery dataset from the Dryad digital repository. From the comparative analysis, DKNN-FL with GLCOA gives Area under the Receiver Operating Characteristic Curve (AUROC) of 0.989 and an accuracy of 98.7 %. The proposed model is applied to another diabetic’s prediction dataset and achieves an accuracy of 95.8 %.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploiting the Intrinsic Neighborhood Semantic Structure for Domain Adaptation in EEG-based Emotion Recognition

引用

IEEE Transactions on Affective Computing 2025年

作者： Yang, Yi Wang, Ze Song, Yu Jia, Ziyu Wang, Boyu Jung, Tzyy-Ping Wan, Feng Macau University of Science and Technology Macao Centre for Mathematical Sciences Respiratory Disease AI Laboratory on Epidemic Intelligence and Medical Big Data Instrument Applications Faculty of Innovation Engineering 999078 China Tianjin University of Technology School of Electrical Engineering and Automation Tianjin Key Laboratory of New Energy Power Conversion Transmission and Intelligent Control Tianjin300384 China Chinese Academy of Sciences Beijing Key Laboratory of Brainnetome and Brain-Computer Interface and Brainnetome Center Institute of Automation Beijing100045 China Western University Department of Computer Science Brain Mind Institute LondonONN6A 3K7 Canada University of California at San Diego Swartz Center for Computational Neuroscience Institute for Neural Computation La Jolla CA92093 United States University of Macau Department of Electrical and Computer Engineering Faculty of Science and Technology China University of Macau Centre for Cognitive and Brain Sciences Centre for Artificial Intelligence and Robotics Institute of Collaborative Innovation 999078 China

Due to the inherent non-stationarity and individual differences present in electroencephalogram (EEG) signals, developing a generalizable model that performs well on new subjects is challenging in EEG-based emotion recognition. Most existing domain adaptation (DA) methods typically mitigate these discrepancies by aligning the marginal distributions of domain feature representations. However, when there is a significant difference in the class-conditional distribution between domain features and labels, the domain-invariant features learned by aligning marginal distributions may have limited discriminative ability for unlabeled target instances or even prove counterproductive. To address this issue, we propose a Neighborhood Semantic Aware Learning-based Dynamic Graph Attention Convolution (NSAL-DGAT) approach that learns target semantic information by considering the inter-domain semantic topological structure, thereby improving classifier adaptation for target instances. Specifically, the proposed NSAL framework is designed to capitalize on the insight that after domain feature alignment, some target samples and their neighboring source samples exhibit similar semantics. By leveraging the neighborhood topological structure, we extract and incorporate semantic target features to train a more transferable classifier. Besides, we implement an entropy weighting mechanism to emphasize representative target semantic information, encouraging target instances to prioritize high-confidence individuals within the source neighborhood. We have conducted extensive experiments on the public SEED dataset and our collected the Hearing-Impaired EEG dataset (HIED). The experimental results underscore the efficacy of our proposed NSAL-DGAT approach, showcasing state-of-the-art accuracy in subject-dependent as well as subject-independent scenarios. The source code is available at https://***/YYingDL/NSAL-DGAT. © 2010-2012 IEEE.

关键词： Speech analysis

来源：评论

学校读者我要写书评

暂无评论

ETiSeg-Net: edge-aware self attention to enhance tissue segmentation in histopathological images

引用

Multimedia Tools and applications 2024年 1-21页

作者： Rashmi, R. Girisha, S. Department of Computer Science and Engineering Manipal Institute of Technology Bengaluru Manipal Academy of Higher Education Manipal India Department of Data Science and Computer Applications Manipal Institute of Technology Manipal Academy of Higher Education Manipal India

Digital pathology employing Whole Slide Images (WSIs) plays a pivotal role in cancer detection. Nevertheless, the manual examination of WSIs for the identification of various tissue regions presents formidable challenges due to its labor-intensive nature and subjective interpretation. Convolutional Neural Network (CNN) based semantic segmentation algorithms have emerged as valuable tools for assisting in this task by automating ROI delineation. The incorporation of attention modules and carefully designed loss functions has shown promise in further augmenting the performance of these algorithms. However, there exists a notable gap in research regarding the utilization of attention modules specifically for tissue segmentation, thereby constraining our comprehension and application of these modules in this context. This study introduces ETiSeg-Net (Edge-aware self attention to enhance Tissue Segmentation), a CNN-based semantic segmentation model that uses a novel edge-based attention module to achieve effective delineation of class boundaries. In addition, an innovative iterative training strategy is devised to efficiently optimize the model parameters. The study also conducts a comprehensive investigation into the impact of attention modules and loss functions on the efficacy of semantic segmentation models. Qualitative and quantitative evaluations of these semantic segmentation models are conducted using publicly available datasets. The findings underscore the potential of attention modules in enhancing the accuracy and effectiveness of tissue semantic segmentation. © The Author(s) 2024.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Flood extent mapping in SAR images using semi-supervised approach

Results in Engineering

引用

Results in Engineering 2025年 26卷

作者： Girisha S Savitha G Sughosh P Department of Data Science and Computer Applications Manipal Institute of Technology Manipal Academy of Higher Education Manipal 576104 Karnataka India Department of Civil Engineering Manipal Institute of Technology Manipal Academy of Higher Education Manipal 576104 Karnataka India

Floods pose a significant threat to both human populations and critical infrastructure. They are caused by excessive precipitation, snowmelt, or infrastructure failures. Precise mapping of flood levels is essential for directing emergency response, allocating resources as efficiently as possible, and estimating the degree of damage. Accurate semantic segmentation of flood-zones from SAR images is necessary for disaster management as it reduces human error and facilitates fast decision-making. However, the complexity of SAR images and the scarcity of annotated datasets make it difficult to develop efficient computer vision algorithms for this use. The paper employs consistency regularization and pseudo-label generation to overcome these issues and presents a novel hybrid semi-supervised semantic segmentation method. In particular, the study presents a student-teacher paradigm in which robust feature learning is ensured by training the teacher model through consistency regularization and self-supervised learning. An attention module-based tailored student model is developed to precisely identify flood-zones. Auxiliary decoders are also added throughout the training phase to reinforce consistency regularization and help the model capture extensive hierarchical characteristics. The suggested method was assessed with publicly accessible datasets, and the outcomes show notable gains in both qualitative and quantitative measures. The findings highlight the method's effectiveness in improving flood-zone segmentation accuracy, leading to more efficient disaster management.

关键词： SAR Semantic segmentation Deep learning Semi-supervised learning Consistency regularization Pseudo-label generation

来源：评论

学校读者我要写书评

暂无评论

Character Classification and Actor Recognition in Yakshagana Images Using Machine Learning Techniques and Facial Makeup Pattern Analysis 8

Character Classification and Actor Recognition in Yakshagana...

引用

8th IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics, DISCOVER 2024

作者： Murthy, Anantha Prathwini Kulkarni, Sanjeev Savitha, G. Nitte Karkala Institute of Computer Science and Information Science Srinivas University Department of Master of Computer Applications India Department of Master of Computer Applications Karkala India Srinivas University Institute of Engineering and Technology Department of Computer Science and Engineering Mangalore India Manipal Institute of Technology Manipal Academy of higher Education Manipal Department of Data Science and Computer Applications India

ISBN: (纸本)9798350350593

Yakshagana, a traditional theater form from Karnataka, India, features a unique combination of vibrant costumes, dynamic dance movements, and elaborate facial makeup, making character and actor identification a challenging task for automated systems. This research work presents a novel approach to classify Yakshagana characters into two primary categories, Vaishnava and Shaiva, and to identify the actors performing these characters using advanced machine learning techniques. Our research employs a Cyclic Gated Recurrent Neural Network (Cyclic GRNN) for classification and identification tasks. For character category classification, we integrate YOLOv3 and Faster R-CNN models, achieving an accuracy of 92.85% with YOLOv3 and 88.285% with Faster R-CNN. The categorization is specifically focused on distinguishing between Vaishnava and Shaiva characters. Additionally, for actor name identification, we utilize a RESNET-50 model, attaining a high accuracy of 95.60%. The results demonstrate the efficacy of Cyclic GRNN combined with state-of-the-art object detection and image classification models in accurately recognizing and categorizing Yakshagana characters and actors. This research contributes to the preservation and digital documentation of Yakshagana by providing robust tools for automated identification, thereby facilitating cultural heritage studies and enhancing audience engagement with this traditional art form. © 2024 IEEE.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Secure Online Banking Authentication System Using Time Bound Password 5

Secure Online Banking Authentication System Using Time Bound...

引用

5th IEEE International Conference on Computing, Power, and Communication Technologies, IC2PCT 2024

作者： Danthy, Ronson Pratham Pai, K. Rao, Vidya Manipal Institute of Technology Manipal Academy of Higher Education Department of Data Science and Computer Applications Manipal India Manipal Institute of Technology Manipal Academy of Higher Education Department of Computer Science and Computer Applications Manipal India

ISBN: (纸本)9798350383522

The banking system is a key component of financial transactions and economic expansion in the modern world. The introduction of internet banking, however, has created some brand-new difficulties, particularly in providing safe user authentication. This study investigates the flaws in conventional authentication techniques, including passwords and SMS-based One-Time Passwords (OTP), and suggests a creative fix. By using Time-Based One-Time Password (TOTP) created by an authenticator app and encrypting the user credentials with this as key, the proposed solution improves security. By addressing problems including poor password habits, SMS theft, OTP delivery delays, phishing, and social engineering attacks, this method strengthens the security of online banking. The approach provides a thorough framework for building this secure authentication system by detailing TOTP generation and encryption. This system successfully defends against popular attacks such as replay attacks, brute force attacks, rainbow tables, packet sniffing, and random guessing. © 2024 IEEE.

关键词： Advanced Encryption Standard - Galois Counter Mode (AES-GCM) Authentication Encryption Online Banking Time-Based One-Time Password (TOTP) Two-Factor authentication (2FA)

来源：评论

学校读者我要写书评

暂无评论

A Novel Eccentric Intrusion Detection Model Based on Recurrent Neural Networks with Leveraging LSTM

引用

computers, Materials & Continua 2024年第3期78卷 3089-3127页

作者： Navaneetha Krishnan Muthunambu Senthil Prabakaran Balasubramanian Prabhu Kavin Kishore Senthil Siruvangur Kavitha Chinnadurai Jehad Ali Department of Computer Applications Karpagam College of EngineeringCoimbatore641032India Department of Computer Science and Engineering Karpagam College of EngineeringCoimbatore641032India Department of Data Science and Business Systems College of Engineering and TechnologySRM Institute of Science and TechnologyKattankulathur603203India Department of Computer Science and Engineering Sathyabama Institute of Science and TechnologyJeppiaar NagarRajiv Gandhi SalaiChennai600119India Department of AI Convergence Network Ajou UniversitySuwon16499Korea

The extensive utilization of the Internet in everyday life can be attributed to the substantial accessibility of online services and the growing significance of the data transmitted via the ***,this development has expanded the potential targets that hackers might *** adequate safeguards,data transmitted on the internet is significantly more susceptible to unauthorized access,theft,or *** identification of unauthorised access attempts is a critical component of cybersecurity as it aids in the detection and prevention of malicious *** research paper introduces a novel intrusion detection framework that utilizes Recurrent Neural Networks(RNN)integrated with Long Short-Term Memory(LSTM)*** proposed model can identify various types of cyberattacks,including conventional and distinctive *** networks,a specific kind of feedforward neural networks,possess an intrinsic memory *** Neural Networks(RNNs)incorporating Long Short-Term Memory(LSTM)mechanisms have demonstrated greater capabilities in retaining and utilizing data dependencies over extended *** such as data types,training duration,accuracy,number of false positives,and number of false negatives are among the parameters employed to assess the effectiveness of these models in identifying both common and unusual *** are utilised in conjunction with LSTM to support human analysts in identifying possible intrusion events,hence enhancing their decision-making capabilities.A potential solution to address the limitations of Shallow learning is the introduction of the Eccentric Intrusion Detection *** model utilises Recurrent Neural Networks,specifically exploiting LSTM *** proposed model achieves detection accuracy(99.5%),generalisation(99%),and false-positive rate(0.72%),the parameters findings reveal that it is superior to state-of-the-art techniques.

关键词： Cybersecurity intrusion detection machine learning leveraging long short-term memory(LLSTM) CICIDS2019 dataset innovative cyberattacks

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：