检索结果-内蒙古大学图书馆

Enhancing intrusion detection in IIoT: optimized CNN model with multi-class SMOTE balancing

Neural Computing and Applications 2024年第24期36卷 14643-14659页

作者： Eid, Abdulrahman Mahmoud Soudan, Bassel Nassif, Ali Bou Injadat, MohammadNoor Department of Computer Engineering College of Computing and Informatics University of Sharjah Sharjah United Arab Emirates Department of Data Science and AI Faculty of Information Technology Zarqa University Zarqa Jordan

This work introduces an intrusion detection system (IDS) tailored for industrial internet of things (IIoT) environments based on an optimized convolutional neural network (CNN) model. The model is trained on a dataset that was balanced using a novel multi-class implementation of synthetic minority over-sampling technique (SMOTE) that ensures equal representation of all classes. Additionally, systematic optimization will be used to fine tune the hyperparameters of the CNN model and mitigate the effects of the increased size of the training dataset. Evaluation results will demonstrate substantial improvement in performance when the optimized CNN model is trained on the balanced dataset. The proposed IDS will be evaluated using the IIoT-specific WUSTL-IIOT-2021 dataset, and then its generalization capability will be verified using the non-domain specific UNSW_NB15 dataset. The model’s performance will be evaluated using accuracy, precision, recall, and F1-score metrics. The results will demonstrate that the proposed IDS is highly effective with performance exceeding 99.9% on all performance metrics. The IDS is also highly effective in detecting intrusion for generic IT networks achieving improvements in excess of 30% compared to the default baseline model. The results emphasize the versatility and effectiveness of the proposed IDS model, making it a reliable and adaptable solution for enhancing network security across diverse network environments. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Natural Language Processing (HNLP) for Model Transformation 5

A Hybrid Natural Language Processing (HNLP) for Model Transf...

引用

5th International Conference on Electronics and Sustainable Communication Systems, ICESC 2024

作者： Madhavi, Banka Chaithanya, T. Rama Babu, S. Giri Pace Institute of Technology Sciences Department of Computer Science & Engineering AP Ongole India Pace Institute of Technology Sciences Department of Computer Science & Engineering AP Ongole India

ISBN: (纸本)9798350379945

Model-to-model transformations involve converting one type of the model into another. This can be used in many ways for various reasons, considering model performance with particular deployment platforms. This paper introduces a Hybrid Natural Language Processing (HNLP) model to visualize model-to-model (M2M) transformation. The proposed HNLP combines Sub-word Embedding (FastText), Encoder-Decoder Architecture, and Attention-based Seq2Seq. The proposed approach focused on processing standard text into meaningful information. Subword Embeddings with FastText retain grammatical aspects by demonstrating words as n-grams, which enhances the interpretation of out-of-vocabulary phrases and unusual word structures. The Encoder-Decoder Architecture translates input sequences into symbolic representations that can be used for various NLP tasks. The Attention-based Seq2Seq method extends this approach by constantly focusing on vital regions of the input sequence during decoding, improving the model's capacity to provide accurate and contextually appropriate outputs. By combining these techniques, the HNLP model enhances sequence transformation quality and gives a more interpretable and visual path for understanding how input transformations occur between models. This visualization capacity is crucial for debugging, model improvement, and understanding the inner workings of complicated NLP systems. Experimental results show that the HNLP model performs well in various NLP tasks, demonstrating its promise as a reliable tool for practical applications and theoretical investigations into model-to-model transformations. © 2024 IEEE.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

Detecting Communities Using Network Embedding and Graph Clustering Approach 6th

Detecting Communities Using Network Embedding and Graph Clus...

引用

6th International Conference on Soft Computing and Signal Processing, ICSCSP 2023

作者： Bhattacharya, Riju Nagwani, Naresh Kumar Tripathi, Sarsij Department of Computer Science and Engineering National Institute of Technology Raipur Raipur India Department of Computer Science and Engineering Motilal Nehru National Institute of Technology Allahabad Allahabad India

ISBN: (纸本)9789819984503

Complex network community structure has been shown effective in many fields, including biology, social media, health, and more. Researchers have explored many different techniques for studying complex networks and discovering communities within them. Most of them, however, lack the expressiveness necessary to learn the node and edge representations of complicated networks. This study has aimed to improve the performance of the Hierarchical Clustering algorithm by analyzing a framework for learning continuous feature representations using node embedding methods for nodes in networks. The proposed method improves upon previous methods by training a map from nodes to a feature space with fewer dimensions. The proposed approach employs the Hierarchical Clustering method to accomplish community identification in the benchmark networks by determining the level of similarity between any pair of node embeddings. Substantial exploratory research on a variety of real-world social networks has shown the efficacy of the proposed approach compared to existing state-of-the-art community discovery techniques. The proposed method has provided high-accuracy results when applied to graph datasets. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Complex networks

来源：评论

学校读者我要写书评

暂无评论

Analyzing Sentiments: A Comprehensive Study of Roberta-Based Sentiment Analysis on Twitters 4

Analyzing Sentiments: A Comprehensive Study of Roberta-Based...

引用

4th International Conference on Advancement in Electronics and Communication engineering, AECE 2024

作者： Krishnamoorthy, A. Sundhar, K.A. Naveen Kumar, V. Karthik, V. Vellore Institute of Technology School of Computer Science & Engineering Tamil Nadu India Vellore Institute of Technology Department of Computer Science Engineering Tamil Nadu India

ISBN: (数字)9798350364729

ISBN: (纸本)9798350364729

Social media platforms have become essential avenues for individuals to express opinions and emotions across diverse topics such as products, events, and policies, using both text and emojis. Understanding these sentiments is crucial, yet manually analyzing the vast data on platforms like Twitter is impractical. Automated sentiment analysis tools, exemplified by Roberta, offer a solution. Roberta, an evolving language model, employs advanced bidirectional learning to predict sentiment labels accurately. This study focuses on sentiment analysis on Twitter, leveraging Roberta to autonomously categorize tweets into positive, negative, or neutral sentiments. We utilizing natural language processing and machine learning algorithms for dataset preprocessing, incorporating sources like Kaggle and ensuring data integrity. Assessment metrics encompassing accuracy, precision, recall, and F1-score, in addition to confusion metrics, assess model performance. According to our findings, machine learning powered sentiment analysis models are dependable, even in the face of challenges and more so when processing Twitter data sets. The study gives direction on how this kind of model can be of help in analyzing sentiments expressed through social networking sites thus assisting scholars and professional Consultancy in the field. © 2024 IEEE.

关键词： Tweets

来源：评论

学校读者我要写书评

暂无评论

Adversarial Robustness in DeepFake Detection: Enhancing Model Resilience with Defensive Strategies 4

Adversarial Robustness in DeepFake Detection: Enhancing Mode...

引用

4th International Conference on Intelligent Cybernetics technology and Applications, ICICyTA 2024

作者： Pallavi, N. Pallavi, T.P. Bylaiah, Sushma Goutam, R. M S Ramaiah Institute of Technology Department of Computer Science and Engineering Karnataka India Karnataka India Atria Institute of Technology Department of Computer Science and Engineering Karnataka India

ISBN: (纸本)9798331506490

Detection of DeepFakes presents formidable challenges primarily due to adversarial attacks that can dramatically reduce model accuracy. Current detection models are reasonably effective in real scenarios;however, these detectors are vulnerable to adversarial attacks, resulting in misclassifications. The current study examines defensive approaches to strengthen the robustness of DeepFake detection approaches when facing adversarial attacks. Explored several alternatives, including adversarial training, input preprocessing, and model ensembles against a manipulated media dataset. The results demonstrate that the model has predicted image appropriately at average of 86.34% but after the adversarial attacks the prediction of the image was 99.8% inappropriate. although adversarial attacks greatly lower the effectiveness of a model, it is possible to improve the robustness of a model through defense methods, particularly adversarial training in combination with an ensemble method which predicted the image appropriately at an average of 52.06%. However, despite the improvement, the defense methods still fail to protect all adversarial examples. The results illustrate the importance of developing new methods of defense against DeepFake detection in light of future adjustments and variants of an attack. © 2024 IEEE.

关键词： adversarial attack attacks deepfake deeplearning detection system

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Based Automated Lip Reading for Deaf 3

Deep Learning Based Automated Lip Reading for Deaf

引用

3rd International Conference for Advancement in technology, ICONAT 2024

作者： Prathyakshini Prathwini Pratheeksha Hegde, N. Vaishali Rashmi, N. Kumar, Archana Praveen NMAM Institute of Technology Department of Information Science and Engineering Karkala Nitte India NMAM Institute of Technology Department of Master of Computer Applications Karkala Nitte India NMAM Institute of Technology Department of Computer Science and Engineering Karkala Nitte India Manipal Academy of Education Manipal Institute of Technology Department of Computer Science and Engineering India

ISBN: (纸本)9798350354171

Speech recognition systems play an integral role in numerous applications, from virtual assistants to accessibility tools. This paper offers a new viewpoint to speech recognition utilizing computer vision and deep learning techniques. The proposed system is trained on a sizable dataset comprising 700 video clips of individuals uttering predefined words, add up to approximately 3 GB of data. Leveraging TensorFlow and Keras, model architecture is designed incorporating convolutional and dense layers. The training process yielded promising results, with a training accuracy of 95.7% and a validation accuracy of 98.5%, indicative of robust classification performance. The integration of computer vision enriches the system's ability to extract meaningful features from audio-visual inputs, enhancing its overall recognition accuracy. Proposed method demonstrates significant potential for real-time speech recognition applications in different areas, like human-computer interaction and assistive technologies. © 2024 IEEE.

关键词： Assistive technology

来源：评论

学校读者我要写书评

暂无评论

Multilevel Thresholding-based Medical Image Segmentation using Hybrid Particle Cuckoo Swarm Optimization

引用

Recent Advances in computer Science and Communications 2024年第5期17卷 12-23页

作者： Kumar, Dharmendra Solanki, Anil Kumar Ahlawat, Anil Kumar Department of Computer Science & Engineering KIET Group of Institutions NCR Delhi Ghaziabad India Department of Computer Science & Engineering Bundelkhand Institute of Engineering and Technology Jhansi India Affiliated to Dr. APJ Abdul Kalam Technical University Lucknow India

Background: The most important aspect of medical image processing and analysis is image segmentation. Fundamentally, the outcomes of segmentation have an impact on all subsequent image testing methods, including object representation and characterization, measuring of features, and even higher-level procedures. The problem with image segmentation is recognition and perceptual completion while segmenting the image. However, these issues can be resolved by multilevel optimization techniques. However, multilevel thresholding will become more computationally intensive with increasing thresholds. Optimization algorithms can resolve these issues. Therefore, hybrid optimization is used for image segmentation in this research work. Methods: The researchers propose a Multilevel Thresholding-based Segmentation using a Hybrid Optimization approach with an adaptive bilateral filter to resolve the optimization challenges in medical image segmentation. The proposed model utilizes Kapur's entropy as the objective function in the nature-inspired optimization algorithm. Results: The result is evaluated using parameters such as the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Feature Similarity Index (FSIM). The researchers perform result analysis with variable thresholding levels on KAU-BCMD and mini-MIAS datasets. The highest PSNR, SSIM, and FSIM achieved were 31.9672, 0.9501, and 0.9728 respectively. The results of the hybrid model are compared with state-of-the-art models, demonstrating its efficiency. Conclusion: The research concludes that the proposed Multilevel thresholding-based segmentation using a Hybrid Optimization approach effectively solves optimization challenges in medical image segmentation. The results indicate its efficiency compared to existing models. The research work highlights the potential of the proposed hybrid model for improving image processing and analysis in the medical field. © 2024 Bentham Science Publishers.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Vulnerability Analysis on Third Party Application 2

Vulnerability Analysis on Third Party Application

引用

2nd IEEE International Conference on Advances in Information technology, ICAIT 2024

作者： Supreetha, H.H. Rajani Suma, H.C. Divya, K.P. Suhasini Computer Science Engineering Department Acharya Institute of Technology bangalore India Electronics Engineering Department BGSIT Institute of Technology BG Nagar India Computer Science Engineering Department Maharaja Institute of Technology Mysore India Electronics Engineering Department Bangalore Institute of Technology Bangalore India

ISBN: (数字)9798350383867

ISBN: (纸本)9798350383867

Road traffic has been an open challenge in a highly-populated area in the world. India is one of them where traffic management is a crucial problem to solve. Road traffic congestion, traffic rule violation, road accidents are the outcome of limitless traffic. The paper attempts to highlight the Real-Time Traffic Control System (RTTCS) monitoring vehicles, passengers, detecting objects on the road, and license plate detection. This paper proposes an RTTCS based on Machine Learning (ML) model, which takes traffic streaming as an input, monitors the traffic by detecting objects, finds the count of the objects per class, detects license plates and recognizes license plate characters using Easy-OCR. For depth analysis, the RTTCS analysis is performed for a particular location at various timestamps for easy traffic monitoring. The proposed system uses the you Only Look once version-4 (YOLOv4) for object detection (e.g., bicycle, motorbike, cars). The transfer learning concept uses pre-trained convolutional weights to optimize the model. © 2024 IEEE.

关键词： Traffic congestion

来源：评论

学校读者我要写书评

暂无评论

Kookaburra Optimization Algorithm and Feature Aggregated Region based Differential network for Colon Cancer Detection from Histopathological Images 8

Kookaburra Optimization Algorithm and Feature Aggregated Reg...

引用

8th International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud), I-SMAC 2024

作者： Gowthamy, J. Ramesh, S.S. Subashka Rm Institute of Science and Technology Amapuram Department of Computer Science and Engineering Chennai India Srm Institute of Science and Technology Ramapuram Department of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350376425

Colon cancer is a significant health concern, ranking as the third leading cause of cancer-related deaths in men and the second in women. Early detection is crucial in reducing the incidence and mortality rates associated with this malignancy. This paper introduces a novel Feature Aggregated Region-based Differential Network (FA-RDNet) designed for the accurate classification of colon cancer from histopathological images. Our approach leverages Multiple Instance Learning (MIL) to effectively handle large datasets structured as bags of instances, addressing the limitations of conventional MIL techniques. A key innovation of our work is the integration of an adaptive feature enhancement module that bilaterally updates multi-level deep features, allowing for efficient extraction of both low-level details and high-level semantic information. Furthermore, we introduce the Kookaburra Optimization Algorithm (KOA) to fine-tune the parameters of FA-RDNet, utilizing a unique two-phase mathematical modeling process inspired by the hunting behaviors of kookaburras. This enhances the model's optimization and classification accuracy. The effectiveness of our proposed system is validated using the publicly available LC25000 dataset, where it achieved a remarkable accuracy of 94.42%, outperforming existing state-of-the-art methods in colon cancer detection. The high computational efficiency and accuracy of FA-RDNet highlight its potential for practical applications in clinical settings, marking a significant advancement in the field of medical image analysis. © 2024 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Improved Software Fault Prediction Model Based on Optimal Features Set and Threshold Values Using Metaheuristic Approach

引用

SN computer Science 2023年第6期4卷 1-12页

作者： Singh, Manpreet Chhabra, Jitender Kumar Computer Engineering Department National Institute of Technology Kurukshetra India

Software fault prediction models are very important to prioritize software classes for effective testing and efficient use of resources so that the testing process’s time, effort, and cost can be reduced. Fault prediction models can be based on either metrics’ threshold values or machine learning. Code metrics’ threshold-based models are easy to automate and faster than machine learning-based models, which can save significant time in the testing process. ROC, Alves ranking, and VARL are famous threshold value calculation techniques. Out of which ROC is the best threshold calculation technique. This research article proposes a new threshold values calculation technique based on metaheuristics. A genetic algorithm and particle swarm optimizer are used to calculate the threshold values, and the proposed technique is tested on ten open-source object-oriented software datasets and four open-source procedural software datasets. Results show that the metaheuristic-based thresholds give better results than ROC-based thresholds. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： GA Metrics’ threshold PSO ROC Software fault prediction

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：