检索结果-内蒙古大学图书馆

Feature Selection Guided by CVOA Metaheuristic for Deep Neural Networks: Application to Multivariate Time Series Forecasting 18th

Feature Selection Guided by CVOA Metaheuristic for Deep Ne...

引用

Proceedings of the 18th International Conference on Soft Computing Models in Industrial and Environmental Applications, SOCO 2023

作者： Jiménez-Navarro, Manuel Jesús Restrepo-Estrada, Camilo Melgar-García, Laura Gutierrez-Aviles, David Department of Computer Languages and Systems University of Seville Seville Spain Faculty of Economic Sciences University of Antioquia Medellín Colombia Data Science and Big Data Lab Pablo de Olavide University Seville Spain

ISBN: (纸本)9783031425288

In the workflow of machine/deep learning, a common question that data scientists ask before training a model is whether to feed the model with the entire set of input dataset features or just a subset of them. In this scenario, the precise selection of features from the input data has a significant impact on the efficiency of model training and the final performance, especially in datasets with many variables. This decision is typically made before model training and in a static way (i.e., the feature selection is applied, and then the model is trained). This paper presents a new methodology based on metaheuristics to dynamically select features. This approach has been applied to train a deep neural network to solve two problems: univariate and multivariate time series forecasting. The performance of the proposed method has been compared to the random forest model (a well-known machine learning model with a highly efficient internal feature selection method). The proposed method outperformed the random forest model in terms of mean squared error, demonstrating the validity of the approach. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

SNN-PAR: Energy Efficient Pedestrian Attribute Recognition via Spiking Neural Networks

arXiv

引用

arXiv 2024年

作者： Wang, Haiyang Zhu, Qian She, Mowen Li, Yabo Song, Haoyu Xu, Minghe Wang, Xiao School of Computer Science and Technology Anhui University Hefei China Faculty of Data Science City University of Macau China

Artificial neural network based Pedestrian Attribute Recognition (PAR) has been widely studied in recent years, despite many progresses, however, the energy consumption is still high. To address this issue, in this paper, we propose a Spiking Neural Network (SNN) based framework for energy-efficient attribute recognition. Specifically, we first adopt a spiking tokenizer module to transform the given pedestrian image into spiking feature representations. Then, the output will be fed into the spiking Transformer backbone networks for energy-efficient feature extraction. We feed the enhanced spiking features into a set of feed-forward networks for pedestrian attribute recognition. In addition to the widely used binary cross-entropy loss function, we also exploit knowledge distillation from the artificial neural network to the spiking Transformer network for more accurate attribute recognition. Extensive experiments on three widely used PAR benchmark datasets fully validated the effectiveness of our proposed SNN-PAR framework. The source code of this paper will be released on https://***/Event-AHU/OpenPAR. Copyright © 2024, The Authors. All rights reserved.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Sparse Attention Decomposition Applied to Circuit Tracing

arXiv

引用

arXiv 2024年

作者： Franco, Gabriel Crovella, Mark Department of Computer Science Boston University United States Department of Computer Science Faculty of Computing & Data Sciences Boston University United States

Many papers have shown that attention heads work in conjunction with each other to perform complex tasks. It's frequently assumed that communication between attention heads is via the addition of specific features to token residuals. In this work we seek to isolate and identify the features used to effect communication and coordination among attention heads in GPT-2 small. Our key leverage on the problem is to show that these features are very often sparsely coded in the singular vectors of attention head matrices. We characterize the dimensionality and occurrence of these signals across the attention heads in GPT-2 small when used for the Indirect Object Identification (IOI) task. The sparse encoding of signals, as provided by attention head singular vectors, allows for efficient separation of signals from the residual background and straightforward identification of communication paths between attention heads. We explore the effectiveness of this approach by tracing portions of the circuits used in the IOI task. Our traces reveal considerable detail not present in previous studies, shedding light on the nature of redundant paths present in GPT-2. And our traces go beyond previous work by identifying features used to communicate between attention heads when performing IOI. © 2024, CC BY.

关键词： Signal encoding

来源：评论

学校读者我要写书评

暂无评论

A trichotomy for transductive online learning 23

A trichotomy for transductive online learning

引用

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Steve Hanneke Shay Moran Jonathan Shafer Department of Computer Science Purdue University Faculty of Mathematics Faculty of Computer Science and Faculty of Data and Decision Sciences Technion - Israel Institute of Technology Computer Science Division UC Berkeley

We present new upper and lower bounds on the number of learner mistakes in the 'transductive' online learning setting of Ben-David, Kushilevitz and Mansour (1997). This setting is similar to standard online learning, except that the adversary fixes a sequence of instances x1,. .., xn to be labeled at the start of the game, and this sequence is known to the learner. Qualitatively, we prove a trichotomy, stating that the minimal number of mistakes made by the learner as n grows can take only one of precisely three possible values: n, Θ(log(n)), or Θ(1). Furthermore, this behavior is determined by a combination of the VC dimension and the Littlestone dimension. Quantitatively, we show a variety of bounds relating the number of mistakes to well-known combinatorial dimensions. In particular, we improve the known lower bound on the constant in the Θ(1) case from Ω(√log(d)) to Ω(log(d)) where d is the Littlestone dimension. Finally, we extend our results to cover multiclass classification and the agnostic setting.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-modal classification of breast cancer lesions in Digital Mammography and contrast enhanced spectral mammography images

引用

computers in Biology and Medicine 2024年 183卷 109266-109266页

作者： Bouzarjomehri, Narjes Barzegar, Mohammad Rostami, Habib Keshavarz, Ahmad Asghari, Ahmad Navid Azad, Saeed Talatian Department of Computer Engineering Faculty of Intelligent Systems Engineering and Data Science Persian Gulf University Bushehr*** Iran Department of Electrical Engineering Faculty of Intelligent Systems Engineering and Data Science Persian Gulf University Bushehr*** Iran

Breast cancer ranks as the second most prevalent cancer in women, recognized as one of the most dangerous types of cancer, and is on the rise globally. Regular screenings are essential for early-stage treatment. Digital mammography (DM) is the most recognized and widely used technique for breast cancer screening. Contrast-Enhanced Spectral Mammography (CESM or CM) is used in conjunction with DM to detect and identify hidden abnormalities, particularly in dense breast tissue where DM alone might not be as effective. In this work, we explore the effectiveness of each modality (CM, DM, or both) in detecting breast cancer lesions using deep learning methods. We introduce an architecture for detecting and classifying breast cancer lesions in DM and CM images in Craniocaudal (CC) and Mediolateral Oblique (MLO) views. The proposed architecture (JointNet) consists of a convolution module for extracting local features, a transformer module for extracting long-range features, and a feature fusion layer to fuse the local features, global features, and global features weighted based on the local ones. This significantly enhances the accuracy of classifying DM and CM images into normal or abnormal categories and lesion classification into benign or malignant. Using our architecture as a backbone, three lesion classification pipelines are introduced that utilize attention mechanisms focused on lesion shape, texture, and overall breast texture, examining the critical features for effective lesion classification. The results demonstrate that our proposed methods outperform their components in classifying images as normal or abnormal and mitigate the limitations of independently using the transformer module or the convolution module. An ensemble model is also introduced to explore the effect of each modality and each view to increase our baseline architecture's accuracy. The results demonstrate superior performance compared with other similar works. The best performance on DM images

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

A Security Enhanced Chaotic-Map-Based Authentication Protocol for Internet of Drones

引用

IEEE Internet of Things Journal 2024年第12期11卷 22301-22309页

作者： Mahmood, Khalid Ghaffar, Zahid Farooq, Muhammad Yahya, Khalid Das, Ashok Kumar Chaudhry, Shehzad Ashraf National Yunlin University Of Science And Technology Graduate School Of Intelligent Data Science Douliu64002 Taiwan National Yunlin University Of Science And Technology Graduate School Of Engineering Science And Technology Douliu64002 Taiwan Abu Dhabi University Department Of Mathematics And Statistics College Of Arts And Sciences Abu Dhabi United Arab Emirates Istanbul Gelisim University Department Of Electrical And Electronics Engineering Istanbul34310 Turkey Sabratah University Department Of Electrical Engineering Faculty Of Engineering Regdalen Libya Center For Security Theory And Algorithmic Research Hyderabad500 032 India Abu Dhabi University Department Of Computer Science And Information Technology College Of Engineering Abu Dhabi United Arab Emirates Nisantasi University Faculty Of Engineering And Architecture Department Of Software Engineering Istanbul34398 Turkey

The Internet of Drones (IoD) extends the capabilities of unmanned aerial vehicles, enabling them to participate in a connected network. In IoD infrastructure, drones communicate not only among themselves but also with users and a control center. This interconnected communication framework holds promise for various applications, from collaborative decision-making to real-time data exchange. However, the expansion of communication in IoD also introduces new challenges, particularly in terms of security, privacy and authentication. Unfortunately, the current authentication protocols are inadequate in offering robust security features against various attacks in the IoD environment. To address these security issues and limitations, we proposed a secure protocol for the IoD environment using chaotic maps and hash functions. In addition, we also employed a physically unclonable function in the development of the proposed protocol. We assess the security of the protocol through both informal and formal security analysis. The formal security analysis is conducted through a widely used random or real (RoR) model. The informal analysis shows the rigorous security features against various attacks, such as masquerading, anonymity violation, and physical cloning attacks. Moreover, we compare the performance of the devised protocol with similar existing protocols across important performance parameters, such as communication overhead, computation overhead, and security features. The devised protocol provides a 67.86% and 17.80% reduction in computation and communication overheads, respectively, as compared to related protocols. The analysis demonstrates the proposed protocol's capacity to support secure communication in the IoD environment and satisfy desirable security attributes. © 2014 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection in the Internet of Things Using Fusion of GRU-LSTM Deep Learning Model

引用

Intelligent Automation & Soft Computing 2023年第8期37卷 2279-2290页

作者： Mohammad S.Al-kahtani Zahid Mehmood Tariq Sadad Islam Zada Gauhar Ali Mohammed ElAffendi Department of Computer Engineering Prince Sattam Bin Abdulaziz UniversityAl-Kharj16273Saudi Arabia Department of Computer Engineering University of Engineering and TechnologyTaxila47050Pakistan The FAMLIR Group The University of LahoreLahore54000Pakistan Department of Computer Science University of Engineering&TechnologyMardan23200Pakistan Faculty of Computing International Islamic UniversityIslamabad44000Pakistan EIAS Data Science and Blockchain Lab College of Computer and Information SciencesPrince Sultan UniversityRiyadh11586Saudi Arabia

Cybersecurity threats are increasing rapidly as hackers use advanced *** a result,cybersecurity has now a significant factor in protecting organizational *** detection systems(IDSs)are used in networks to flag serious issues during network management,including identifying malicious traffic,which is a *** remains an open contest over how to learn features in IDS since current approaches use deep learning *** learning,which combines swarm intelligence and evolution,is gaining attention for further improvement against cyber *** this study,we employed a PSO-GA(fusion of particle swarm optimization(PSO)and genetic algorithm(GA))for feature selection on the CICIDS-2017 *** achieve better accuracy,we proposed a hybrid model called LSTM-GRU of deep learning that fused the GRU(gated recurrent unit)and LSTM(long short-term memory).The results show considerable improvement,detecting several network attacks with 98.86%accuracy.A comparative study with other current methods confirms the efficacy of our proposed IDS scheme.

关键词： Cyber security deep learning intrusion detection PSO-GA CICIDS-2017 intelligent system security and privacy IoT

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Communication with Deep Learning and XAI for Enhanced Arabic Sign Language Recognition

arXiv

引用

arXiv 2025年

作者： Balat, Mazen Awaad, Rewaa Zaky, Ahmed B. Aly, Salah A. CS & IT Department Egypt-Japanese University of Science & Technology Alexandria Egypt Faculty of Computing and Data Science Badya University Giza Egypt Computer Science Section Faculty of Science Fayoum University Fayoum Egypt

This study introduces an integrated approach to recognizing Arabic Sign Language (ArSL) using state-of-the-art deep learning models such as MobileNetV3, ResNet50, and EfficientNet-B2. These models are further enhanced by explainable AI (XAI) techniques to boost interpretability. The ArSL2018 and RGB Arabic Alphabets Sign Language (AASL) datasets are employed, with EfficientNet-B2 achieving peak accuracies of 99.48% and 98.99%, respectively. Key innovations include sophisticated data augmentation methods to mitigate class imbalance, implementation of stratified 5-fold cross-validation for better generalization, and the use of Grad-CAM for clear model decision transparency. The proposed system not only sets new benchmarks in recognition accuracy but also emphasizes interpretability, making it suitable for applications in healthcare, education, and inclusive communication technologies. © 2025, CC0.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

E-Learning: A Successful Learning Environment 1

E-Learning: A Successful Learning Environment

引用

1st International Conference on computer science and Emerging Technologies, CSET 2023

作者： Momani, Fayiz Hatamleh, Amer Ahmad Amir Fazamin Wan Hamzah, Wan Mohd Jusoh, Julaily Aida Zakaria, Muhammad D. Das, Sunanda Faculty of Management and Financial Sciences E-business & Commerce Department Jordan MIS Bi and Data Analysis Irbid National University Faculty of Business and Financial Science Irbid Jordan Faculty of Informatics and Computing UniSZA Malaysia School of Computer Science and Engineering Jain University Cyber Security Bangalore India

ISBN: (纸本)9798350341737

Education is developing with the economy. This led to the creation of e-Iearning tools for instructors and students. In e-learning, electronic resources are used to teach. E-learning can be taught in or out of school, but computers and the Internet are essential. That is, e-learning involves delivering education to a big group of people at once or over time. So, overall, in this research is about E-learning that make a successful environment. We chose this term since it is widely used in E-learning research to improve student academic achievement. Due to the difficulties of accessing adequate study materials during the Covid-19 epidemic, distance learning has grown more popular among students. This study proposes a unique learning technique that involves choosing and organizing important learning items using a recommender system. We also studied our method's efficacy. Using a recommended system to help online learning activities was proven to be effective. Software development methodology is a collection of principles and processes used to build software. Systematic software development is the goal. The research approach utilized for this topic is Lean Development Methodology. We also examine the advantages and disadvantages of e-learning. This research also introduces various new strategies and methodologies that may be employed in the future to improve system-learning. It also presents system that we choose and why we choose it. So, it aims to eliminate waste and increase production. Following the principles can help developers avoid nonproductive chores while still producing high-quality results. The ultimate objective is to develop an efficient and error-free system. Finally, E-learning is a social revolution. It's part of a bigger rethinking of how we teach future employees and students. The eLearning platform enabled the learning data visualization dashboard beautifully. User-centered design lets the dashboard adapt to its users' demands. © 2023 IEEE.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Detecting Food Items with Possible Allergens Using Deep Neural Network

Detecting Food Items with Possible Allergens Using Deep Neur...

引用

Cybernetics and Innovations (ICCI), International Conference on

作者： Kannapak Yudeepanao Kittiphop Sakkato Pornthanwa Tosaman Pokpong Songmuang Rachada Kongkachandra Data Science and Innovation College of Interdisciplinary Studies Thammasat University Thailand Department of Computer Science Faculty of Science and Technology Thammasat University Thailand

ISBN: (数字)9798350370096

ISBN: (纸本)9798350370102

Food allergies are a significant concern for the community as they can have adverse effects on human health. Even a small amount of certain food items can trigger allergic reactions in the body, ranging from mild symptoms like hives or itchiness to severe and life-threatening anaphylaxis. However, these reactions can often be prevented by being aware of the specific allergen-based food items and avoiding their consumption. Our research aims to address this issue by utilizing a Convolutional Neural Network-based object. This model, belonging to the realm of deep learning techniques, is utilized to identify food items present in images. The primary objective behind this recognition process is to facilitate the development of a system that can effectively identify ingredients capable of triggering allergic reactions. This study explored three deep learning techniques for the classification of food images. The effectiveness of three distinct deep learning models, namely MobilenetV2, EfficientNetB0 and Resnet50v2, was examined. Notably, the EfficientNetB0 classifier outperformed the others, achieving an impressive accuracy score of 97%.

关键词： Deep learning Privacy Adaptation models Technological innovation Transfer learning data models Safety

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：