检索结果-内蒙古大学图书馆

2024 Symposium on Beyond the Canopy: Technologies and Applications of Remote Sensing

作者： Hnatushenko, Viktoriia Hnatushenko, Volodymyr Soldatenko, Dmytro Dept. Information Technologies and Systems Ukrainian State University of Science and Technologies Dnipro Ukraine Dept. Information Technologies and Computer Engineering Dnipro University of Technology Dnipro Ukraine Institute of Photogrammetry and GeoInformation Leibniz Universität Hannover Germany

Today's agricultural sector is characterized by an important role of accurate mapping and monitoring of agriculture with the help of satellite imagery, which allows to optimize the use of resources, to plan crop areas and to forecast productivity. Classification of satellite images with unbalanced sample distribution is a critical problem in this regard. Traditional machine learning algorithms in particular have limitations in dealing with sample imbalance. In this paper, we proposed convolution neural networks for semantic segmentation, where sample imbalance is considered based on a particular loss function coupled with data augmentation. To illustrate our method, we use Sentinel-2 remote sensing (RS) images covering a number of regions in Ukraine, and then we create an image dataset of the region and for training and testing make data augmentation. The models with different architectural features were investigated. The results demonstrate that the proposed CNN has a higher classification accuracy than the ones discussed in the paper: the classification accuracy on the test dataset reached 96.7% with intersection-over-union values of up to 89.7%. This opens the way for further research in the direction of refining algorithms for classify satellite data with an imbalanced class structure. © 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

Relation-based Counterfactual data Augmentation and Contrastive Learning for Robustifying Natural Language Inference Models

arXiv

引用

arXiv 2024年

作者： Yang, Heerin Hwang, Seung-Won So, Jungmin Dept. of Computer Science and Engineering Sogang University Korea Republic of Dept. of Computer Science and Engineering Seoul National University Korea Republic of LG Electronics Korea Republic of

Although pre-trained language models show good performance on various natural language processing tasks, they often rely on non-causal features and patterns to determine the outcome. For natural language inference tasks, previous results have shown that even a model trained on a large number of data fails to perform well on counterfactually revised data, indicating that the model is not robustly learning the semantics of the classes. In this paper, we propose a method in which we use token-based and sentence-based augmentation methods to generate counterfactual sentence pairs that belong to each class, and apply contrastive learning to help the model learn the difference between sentence pairs of different classes with similar contexts. Evaluation results with counterfactually-revised dataset and general NLI datasets show that the proposed method can improve the performance and robustness of the NLI model. © 2024, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Comparative Evaluation of Machine Learning-Based Controller Area Network Intrusion Detection Systems 14

Comparative Evaluation of Machine Learning-Based Controller ...

引用

14th International Conference on Information and Communication Systems, ICICS 2023

作者： Al-Jarrah, Omar Y. Ramadan, Omar A. Alsaify, Baha Alkhushayni, Suboh Jordan University of Science and Technology Faculty of Computer and Information Technology Dept. Network Engineering and Security P.O. Box 3030 Irbid22110 Jordan Jordan University of Science and Technology Faculty of Engineering Dept. Mechanical Engineering P.O. Box 3030 Irbid22110 Jordan Yarmouk University Dept. Information Systems Irbid Jordan

ISBN: (纸本)9798350307870

The controller area network (CAN) protocol is widely used in vehicle networks. However, it lacks essential security features like confidentiality and authentication. To enhance vehicle security, researchers have proposed and developed various methods, including intrusion detection systems (IDSs). These IDSs play a key role in early detection of potential cyberattacks, safeguarding vehicles, passengers, and road users. Machine learning (ML) algorithms are increasingly employed in CAN IDSs but they face challenges due to diverse experimental settings, datasets, and metrics used for evaluation, impeding progress and comparison of different approaches in this field. This study aims to benchmark well-recognized ML algorithms, including gaussianNB, k-nearest neighbors (KNN), decision trees (DT), random forest (RF), long short-term memory (LSTM), and convolutional neural networks (CNN), for intrusion detection in CAN. We evaluate the performance of these algorithms using seven evaluation metrics on the real ORNL automotive dynamometer (ROAD) dataset, a state-of-the-art benchmark in this field. Experimental results demonstrate that KNN, DT, RF, LSTM, and CNN can detect various types of attacks (e.g., Fuzzing, targeted ID, and masquerade attacks) with high accuracy. However, the computational and time efficiency of tree-based classifiers, namely DT and RF, makes them a potentially appealing choice for real-time intrusion detection tasks in CAN. © 2023 IEEE.

关键词： Intrusion detection

来源：评论

学校读者我要写书评

暂无评论

Rule-Augmented Neural Networks for Improved Answer Validation in Question Answering Systems

Rule-Augmented Neural Networks for Improved Answer Validatio...

引用

Computational Intelligence Approaches and Applications (ICCIAA), International Conference on

作者： Eman Alduweib Waseem Alromema Nisreen Otoum dept. of Computer Science University of Petra Amman Jordan dept. of Computer Science and Information Taibah University Medinah Saudi Arabia dept. of Software Engineering University of Petra Amman Jordan

ISBN: (数字)9798331523657

ISBN: (纸本)9798331523664

This work presents a hybrid method for answer validation in question answering systems, combining neural network methods with explicit logical rules. The Rule-Augmented Neural Network (RANN) framework integrates the expressiveness of neural networks and precision of rule-based systems for improving answer validation accuracy. Using a sample of the Natural Questions dataset, we demonstrate how we can achieve more reliable and interpretable answer validation by complementing traditional neural architectures with domain-specific rules. The experimental results validate the RANN framework. The experimental results emphasize that the hybrid approach outperforms both neural-based models and rule-based reasoning by achieving the best accuracy, precision, recall, and F1 score.

关键词： Accuracy Computational modeling Neural networks computer architecture Question answering (information retrieval) Cognition Reliability Artificial intelligence Computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Mobile Charger Scheduling Algorithm for Energy Recharging in Wireless Sensor Networks

Mobile Charger Scheduling Algorithm for Energy Recharging in...

引用

2022 International Conference on Electronic Systems and Intelligent Computing, ICESIC 2022

作者： Dande, Bhargavi Chang, Chih-Yung Fan, Cheng-De Tamkang University Dept. of Computer Science and Information Engineering New Taipei City Taiwan Tamkang University Dept. of Electrical and Computer Engineering New Taipei City Taiwan

ISBN: (纸本)9781665483858

In wireless sensor networks (WSNs), the finite battery capacity of the sensor nodes restricts the lifetime of the network. With the wireless power transfer (WPT) technology, the mobile chargers can transmit energy to the sensor nodes, which provides a solution for recharging the sensors. Therefore, this paper proposed a mobile charger scheduling (MCS) algorithm, which maximizes the coverage quality of the network. The proposed MCS algorithm formulated an adaptive recharging request threshold for the sensor nodes. Then the coverage benefit (CB) of the requested sensors is considered to dynamically adjust the path. Finally, through the experimental analysis, it is shown that the proposed MCS algorithm gains the best performance than the existing works. © 2022 IEEE.

关键词： Scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

An Efficient Authentication and Key Agreement for Vehicular communications 3

An Efficient Authentication and Key Agreement for Vehicular ...

引用

3rd IEEE International Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation, IATMSI 2025

作者： Rajasekaran, Arun Sekar Obulesu, A. Maria, Azees Sivakumar, T. Vimalnath, S. Sr University Dept. of Ece Telangana Warangal 506371 India Vidya Jyoti Institute of Technology (A) Dept of Information Technology Hyderabad India VIT-AP University School of Computer Science and Engineering Amaravati India Sri Eshwar College of Engineering Department of Computer Science and Engineering Coimbatore India M.Kumarasamy College of Engineering Dept. of Ece Thalavapalayam Karur 639113 India

ISBN: (纸本)9798331521691

Vehicular Ad Hoc Networks (VANETs) are crucial in improving transportation infrastructure by supporting vehicle to infrastructure (V2I) communication. Nevertheless, maintaining secure and efficient interoperability in such environments has always been a big push because of high mobility, and the vulnerability to cyber threats. This paper puts forward a novel authentication and key agreement protocol to be used in V2I communication settings of VANETs taking into consideration security and efficiency factors. The work in the proposal uses a lightweight cryptographic solution to maintain authentication of vehicles and the infrastructure without incurring much computational and communication load. In general, it is possible to use elliptic curve cryptography (ECC), and hash functions to protect the exchange from basic threats like impersonation, replay and so on. In addition, the proposed protocol includes a session key agreement mechanism that builds the secure communication channels in a dynamic way and provides confidentiality and integrity for the data during the exchange of messages. The simulation results show that the proposed scheme has better performance and it has lower computational overhead and communication overhead than the other related protocols. © 2025 IEEE.

关键词： Authentication digital signature privacy security VANET

来源：评论

学校读者我要写书评

暂无评论

Investigating the Efficacy of Diverse Machine Learning Classifiers for Parkinson’s Disease Detection

Investigating the Efficacy of Diverse Machine Learning Class...

引用

International Conference on Computing and Networking Technology (ICCNT)

作者： Sourabarna Roy Tina Debbarma Tannistha Pal Prakash Duraisamy Dept. of Computer Science and Engineering National Institute of Technology Agartala Agartala India Dept. of Computer Science and Engineering University of Wisconsin- Green Bay WI United States

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

To detect Parkinson’s disease, we compare the effectiveness of K-Nearest Neighbors (KNN), Logistic Regression (LR), Support Vector Machines (SVM), and Random Forest (RF) algorithms. Utilizing a dataset with clinical and biomedical features, we preprocess the data to handle missing values and standardize the features. Subsequently, we train each algorithm with the preprocessed data and evaluate their performance using metrics like accuracy, precision, recall, and F1-score. Our results indicate that all four algorithms achieve excellent accuracy in diagnosing Parkinson’s disease, with KNN slightly outperforming the others. However, the selection of the algorithm may depend on specific needs such as interpretability and computational efficiency. Additionally, we conduct a feature importance analysis to identify the most relevant features for Parkinson’s disease identification, offering insights that can aid in early diagnosis and disease management.

关键词： Support vector machines Radio frequency Accuracy Sensitivity Machine learning algorithms Nearest neighbor methods Sensitivity and specificity Vectors Random forests Diseases

来源：评论

学校读者我要写书评

暂无评论

Discriminative Vector Learning with Application to Single Channel Speech Separation 48

Discriminative Vector Learning with Application to Single Ch...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Tan, Ha Minh Liang, Kai-Wen Wang, Jia-Ching National Central University Dept. of Computer Science and Information Engineering Taiwan Nguyen Tat Thanh University Faculty of Information Technology Viet Nam

ISBN: (纸本)9781728163277

In this paper, we introduce a discriminative vector learning method and apply it to single-channel speech separation. First, speech samples are transformed into discriminative vectors using two backbone networks. These vectors are easily separated by simple clustering algorithms. Among them, vectors with lower similarity are separated into different clusters, while vectors in the same cluster have higher similarity. This property is very important in image segmentation, audio separation, and data clustering problems. In our work, we design the network architecture to improve the discriminativeness of vectors through learning, taking this task as spectrogram segmentation. Experiments show that our method significantly improves performance compared to other deep clustering methods for speech separation. © 2023 IEEE.

关键词： Vectors

来源：评论

学校读者我要写书评

暂无评论

Ethical Considerations in data Mining and Analytics 1

Ethical Considerations in Data Mining and Analytics

引用

1st IEEE International Conference on Advances in Computing, Communication and Networking, ICAC2N 2024

作者： Kaushik, Priyanka Sharma, Khushboo Mahawar, Mukesh Kumar Wasim, Javed Dey, Gunadhar Nibiya, S. Anto Chandigarh University Dept. of Com. Science Engineering Chandigarh India Madhav University Dept. of Commerce & Management Pindwara India Mangalayatan University Dept. of Computer Eng. & Applications Aligarh India Usha Martin University Dept. of Computing & IT Jharkhand India Meenakshi A.H. Edu. & Research Dept. of Mathematics Chennai India

ISBN: (纸本)9798350356816

As data mining & analytics technologies advance, they offer unprecedented opportunities for extracting valuable insights from vast datasets. However, these advancements bring forth a complex array of ethical considerations that must be addressed to ensure responsible practices in the field. This paper explores the ethical dimensions of data mining & analytics, focusing on key issues such as privacy, consent, & bias. Privacy concerns arise from the extensive collection & potential misuse of personal data, necessitating stringent measures to protect individuals' information. Consent, a fundamental ethical principle, requires transparent practices where data subjects are informed & agree to the use of their data. Additionally, the challenge of algorithmic bias highlights how data-driven models can perpetuate or exacerbate existing inequalities if not carefully managed. The paper also examines the ethical implications of data ownership & the responsibilities of organizations in managing & communicating data effectively. Through a review of current practices, case studies, & regulatory frameworks, this study aims to provide a comprehensive overview of the ethical landscape in data mining & analytics. It proposes a set of best practices for professionals to navigate these challenges, emphasizing the need for ethical guidelines, ongoing education, & robust oversight mechanisms. By addressing these concerns, the paper seeks to foster a more responsible & equitable approach to data mining & analytics that balances innovation with the protection of individual rights & societal values. © 2024 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Advanced Feature engineering and Sampling Technique for Effective Heart Disease Prediction System using Machine Learning Technique 5

Advanced Feature Engineering and Sampling Technique for Effe...

引用

5th IEEE International Conference on Electro-Computing Technologies for Humanity, NIGERCON 2024

作者： Awoyemi, John Oloruntoba Awoyemi, Bosede Olanike Ugwu, Tolulope Anthonia Oluwasola, Atolase Mercy Oladokun, Temitope Ugwu, Chukwuemeka Christian Federal University Oye-Ekiti Dept. of Computer Science Oye-Ekiti Nigeria Afe Babalola University Dept. of Economics Ado-Ekiti Nigeria Federal University of Technology Dept. of Information Systems Akure Nigeria International Brewery PLC-AB-Inbev Lagos Nigeria Federal University of Technology Dept. of Data Science Akure Nigeria

ISBN: (纸本)9798331542559

Predicting heart disease is crucial in healthcare, where accurate detection can significantly impact patient outcomes. A major challenge in prediction is imbalanced data, with underrepresented samples in the minority class leading to biased models. This study addresses this by combining advanced feature engineering with SMOTE (Synthetic Minority Oversampling Technique) to manage class imbalance. The objectives include refining data through feature engineering, building predictive models using machine learning algorithms, and assessing performance metrics. Algorithms such as Random Forest, SVM, Decision Tree, Logistic Regression, and K Nearest Neighbor are applied. Results show that robust feature engineering notably enhances performance, with Random Forest achieving the highest accuracy (0.9604), followed by SVM (0.954) and Decision Tree (0.947). The findings highlight the role of feature engineering and sampling in improving heart disease prediction models, with meaningful implications for healthcare advancements. © 2024 IEEE.

关键词： Algorithms Class imbalance Coronary data Mining SMOTE

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：