检索结果-内蒙古大学图书馆

Innovations in Intelligent Systems and applications Conference

作者： Gokhan Cetiner Cem Baglum Ugur Yayan Ahmet Yazici Computer Engineering Department Center of Intelligent Systems Applications Research (CISAR) Eskisehir Osmangazi University Eskisehir Turkiye Software Engineering Department Center of Intelligent Systems Applications Research (CISAR) Eskisehir Osmangazi University Eskisehir Turkiye

ISBN: (数字)9798350379433

ISBN: (纸本)9798350379440

Artificial intelligence is actively becoming widespread in every field. The complex structure of deep neural network models, especially those used in real-world applications, may contain faults that lead to unexpected model behavior, even well-trained AI models. To detect these types of faults and model behaviors, it is necessary to perform a white box test of the model's reliability and adequacy by mutating its layers, activation functions, and similar model structures and parameters. White box testing is a method that tests the logical flow of software to detect possible faults and deficiencies in the code. The proposed module performs model-based testing using this testing technique. The module was developed as an extension of the traditional testing tool IM-FIT. It was inspired by IM-FIT's advanced fault detection capacity in traditional software testing, autonomous or customized testing options, and detailed report generation capabilities. The module carries out an advanced scanning and mutation process using its own unique fault library for deep neural networks. The module successfully performed the test by autonomously generating mutants and classifying these mutants. Tests were performed on a Transformer model using the TensorFlow library during the testing process. As a result of the tests performed with 5, 10, and 20 epoch values, a total of 606 mutants were produced. While 414 of these mutants were classified as killed, 192 were classified as survived. This study presents it to the literature as an effective mutation testing tool focusing on deep neural networks.

关键词： software testing Technological innovation Fault detection Focusing Artificial neural networks Transformers Libraries software reliability Intelligent systems Glass box

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection Mechanism Using Deep Learning

Intrusion Detection Mechanism Using Deep Learning

引用

IoT Based Control Networks and Intelligent Systems (ICICNIS), International Conference on

作者： P. Ananthi K. Nirmaladevi Naveen Kumar S Department of Computer Technology-PG Kongu Engineering College Erode Tamilnadu Department of Computer Applications Madanapalle Institute of Technology & Science M.Sc Software Systems Kongu Engineering College Erode India

ISBN: (数字)9798331518097

ISBN: (纸本)9798331518103

In today's connected and data-driven world, networks and digital systems need to be protected from malicious attacks. The effectiveness of conventional Intrusion Detection Systems (IDS) in recognizing and impeding novel threats and intricate methodologies has its limits. This study introduces a new method for improving network security by using deep learning techniques in the design and development of an intrusion detection system (IDS). The suggested Intrusion Detection System uses deep learning techniques, specifically Long Short Term Memory (LSTM) and Convolutional Neural Networks (CNNs), to identify and react to known and new security threats. Data can be taught to be encoded into a lower-dimensional representation and then decoded back into its original form using autoencoders (AEs). Autoencoders are a cutting-edge intrusion detection system that can lower false positives and raise overall detection accuracy because of their capacity to automatically learn from and adjust to the ever-changing threat landscape. They enhance computer networks capacity to spot irregularities and possible security threats. With the deep learning model obtaining a high accuracy of 99% in real-world threat detection, the results show that deep learning is effective in developing intrusion detection systems.

关键词： Deep learning Accuracy Autoencoders Intrusion detection Network security Threat assessment Real-time systems computer security Optimization Long short term memory

来源：评论

学校读者我要写书评

暂无评论

Video Surveillance-Based Intrusion Detection System in Edge Cloud Environment 3rd

Video Surveillance-Based Intrusion Detection System in Edge ...

引用

3rd International Conference on Emergent Converging Technologies and Biomedical Systems, ETBS 2023

作者： Sharma, Annu Devasenapathy, Deepa Raja, M. Shadrach, Finney Daniel Shirgire, Anil Arun, R. Yau, Thomas Moh Shan Department of Computer Science and Applications Rajarajeshwari College of Engineering Karnataka Bengaluru560074 India Instructor-II/Computing and Software Engineering U.A. Whitaker College of Engineering Florida Gulf Coast University Fort Myers United States Department of Computer Science and Engineering CMR Institute of Technology Karnataka Bengaluru560037 India Department of Electronics and Communication Engineering KPR Institute of Engineering and Technology Tamil Nadu Coimbatore641407 India Department of Civil Engineering Dr. D Y Patil Institute of Technology Maharashtra Pimpri Pune411018 India Department of Computer Science and Engineering Hindusthan Institute of Technology Coimbatore India Director of Centre of University Courses and Innovative Learning University of Technology Sarawak Jalan Universiti Sibu Sarawak96000 Malaysia

ISBN: (纸本)9789819986453

Video surveillance is mostly used for monitoring distant videos and the cloud memory in the system provides more secure to the data. But, in the existing methodologies poor intrusion detection accuracy as well as poor performance rate is considered as the main demerits. So, to overcome the above-mentioned demerits, a novel video surveillance-based Intrusion Detection System (IDS) (VS-IDS) in the edge cloud environment via deep learning method has been developed in this research. Different algorithms are used for object detection and intrusion detection. Subsequently, the proposed model can be implemented and verified in NVIDIA Jetson TX2 platform. The proposed model attains better rate of performance than the existing models. The Performance of the proposed model can be validated with accuracy, precision, recall, MSE and F-measure. Higher performance rate indicates that the system can detect intrusion more accurately than the existing mechanisms. 99% of detection accuracy can be attained through the proposed model. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Intrusion detection

来源：评论

学校读者我要写书评

暂无评论

Data Analytics for Core Temperature Estimation in Battery Management System for Electric and Hybrid Vehicles 5

Data Analytics for Core Temperature Estimation in Battery Ma...

引用

5th International Conference on Energy, Power, and Environment, ICEPE 2023

作者： Chhetri, Ahilya Surya, Sumukh Mohan Krishna, S. Rao, Vidya Cactus Communications Engineering and Technology Mumbai India Bosch Global Software Technologies Pvt. Limited Bengaluru India Aced Alliance University Department of Electrical and Electronics Engineering Bengaluru India Manipal Institute of Technology Manipal Academy of Higher Education Department of Data Science and Computer Applications Manipal India

ISBN: (纸本)9798350313123

Compared to conventional means of transport thatuse fossil fuels, electric vehicles are known to reduce pollution levels as they take power from renewable sources of energy stored in energy storage devices like batteries and fuel cells. Core temperature estimation of batteriesis extremely important inbattery management systems for preventing thermal runaway and ensuring safe operation. In this study, core temperature is estimated using a Kalman filter for two different battery chemistries viz;lithium polymer and lithium iron phosphate using a second-order thermal model. Further, a linear regression model is applied to verify the prediction over the trained and tested dataset. The lithium iron phosphate prediction curvehad a fit of approximately 82-83%, whilelithium polymerhad a fit of 72-82% during the charging and discharging of OCV-SOC variation. © 2023 IEEE.

关键词： Simulink

来源：评论

学校读者我要写书评

暂无评论

Explainable Artificial Intelligence Approaches: Challenges and Perspectives

Explainable Artificial Intelligence Approaches: Challenges a...

引用

2021 IEEE International Conference "Quality Management, Transport and Information Security, Information Technologies", T and QM and IS 2021

作者： Shevskaya, Natalya V. Saint Petersburg Electrotechnical University 'LETI' Software Engineering and Computer Applications Department Saint Petersburg Russia

ISBN: (纸本)9781665433020

Artificial intelligence systems are used in many areas, for example, in finance and medicine. Every year, intelligent systems get for processing more and more data and make more and more decisions. All these decisions are exerting more and more influence on human destinies. Mistrust of fully non-human, autonomous artificial intelligence systems become the cornerstone. The key of mistrust is hidden in the lack of understanding as to why intelligent systems make this or that decision. To solve this problem, methods of explainable artificial intelligence are used. This article provides an overview of such methods and discusses the main trends in their development, as well as the implementation difficulties that one has to face during development and use. © 2021 IEEE.

关键词： Intelligent systems

来源：评论

学校读者我要写书评

暂无评论

Retrieval-Augmented Generation for Multiple-Choice Questions and Answers Generation

引用

Procedia computer Science 2025年 259卷 504-511页

作者： Pradeesh N Remya T MG Thushara K Arun Krishna Pranav V Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India Omnex Software Solutions Pvt. Ltd Thoraipakkam Chennai India Department of Computer Science and Applications Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India

Generating a large volume of diverse questions for educational purposes can be a challenging and time-consuming task for educators. Traditional methods like manually writing questions or using templates often fall short when it comes to variety and relevance. In order to cope with these di ffi culties, we have examined a di ff erent approach adopting Retrieval-Augmented Generation ( RAG). RAG is an enhancement technique that allows retrieving independent documents and combines it with artificial intelligence that is able to qualitatively generate contextually adequate questions. In this research, RAG was employed through the Ample LMS platform in which PDF documents were the source of generating MCQ questions. The results of the research indicate that with this method, the process of formulating the questions is relatively faster and the questions generated are of better and diverse in nature. With RAG, the burden of constructing questions is lessened for teachers and the students are provided with better interactive learning opportunities. This approach presents a better alternative that is flexible and suitable for various educational requirements, demonstrating the e ff ectiveness of AI in the improvement of the learning and teaching processes.

关键词： RAG Education Question Generation LLM

来源：评论

学校读者我要写书评

暂无评论

Explainable artificial intelligence methods based on feature space analysis 4

Explainable artificial intelligence methods based on feature...

引用

4th International Conference on Control in Technical Systems, CTS 2021

作者： Popov, Nikita V. Shevskaya, Natalya V. Department of Software Engineering and Computer Applications Saint Petersburg Electrotechnical University "LETI" St. Petersburg Russia

ISBN: (纸本)9781665424547

In the 21st century, mankind is actively introducing machine learning and artificial intelligence into all spheres of life. But most modern algorithms output the final result of the calculations without revealing the details of obtaining the result, which is the reason for some skepticism towards it. To correct this situation, there is a need to use understandable machine learning methods that increase the transparency of use and the level of trust of people. The work reviews existing solutions to this problem, and also draws a conclusion on the effectiveness of a particular algorithm. Based on the results of the article, ways to further develop the work are proposed. © 2021 IEEE.

关键词： Lime

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Approaches for Efficient Traffic Flow in Smart Cities 3

Machine Learning Approaches for Efficient Traffic Flow in Sm...

引用

3rd Odisha International Conference on Electrical Power engineering, Communication and Computing Technology, ODICON 2024

作者： Kulkarni, Aniket Anitha, P. Valluri, John Yesudas Sunena Rose, M.V. Hemavathi, U. Hussein, Omar M. Software Qa Manager Drata Inc San DiegoCA United States Kalasalingam Academy of Research and Education Department of Computer Applications Tamil Nadu Srivilliputhur India Faculty of Management Studies Bengaluru India Department of Computer Science and Engineering Bengaluru India Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Department of Cse Avadi Chennai India Al-Farahidi University Medical Technical College Medical Instrumentation Engineering Baghdad Iraq

ISBN: (纸本)9798350354379

These days, a lot of towns struggle with traffic jams during peak hours, which increment contamination, commotion, and feelings of anxiety for the general population. Because neural networks (NN) and machine learning (ML) techniques can handle large numbers of parameters in massive amounts of data and dynamic behaviour over time, they are replacing analytical and statistical methods in the solving of real-world problems. This research presents a convergence of machine-learning (ML) and deep-learning (DL) calculations for traffic stream prediction;this paves the way for flexible traffic signals, which can be implemented through the use of a calculation that adjusts the timing in accordance with the anticipated traffic flow or by controlling traffic signals to some degree. The proposed ML and DL models are constructed, validated, and tested using two publicly accessible datasets. The first depicts the total number of cars subjected to checks using various sensors at six junctions over the course of 56 days. In this research, ML and DL models are created using data from four of the six junctions. The excellent performance metrics obtained by all ML and DL computations suggest that their application to intelligent traffic light management is feasible. © 2024 IEEE.

关键词： Traffic congestion

来源：评论

学校读者我要写书评

暂无评论

Blood Cell Classification and Counting in Hematology through Deep Learning

Blood Cell Classification and Counting in Hematology through...

引用

Sustainable Expert Systems (ICSES), International Conference on

作者： P. Ananthi K Nirmala Devi T. E. Ramya P. Shanmugapriya T. Aradhya Department of Computer Technology Kongu Engineering College Perundurai Erode Tamil Nadu Department of Computer Applications Madanapalle Institute of Technology & Science M.Sc Software Systems Kongu Engineering College Perundurai Erode Tamil Nadu

ISBN: (数字)9798331540364

ISBN: (纸本)9798331540371

Medical diagnosis is one of the areas that has been greatly influenced by the progress of computer vision technologies. This study presents a unique approach for the detection and counting of blood cells in hematology: You Only Look One algorithm version 7 (YOLOv7). YOLO is designed for real-time object detection, making it ideal for applications such as blood vessel detection where speed is critical. The architecture allows the model to process images faster than other recognition models, such as R-CNN or SSD, which is very important for situations that require fast results, such as analysis automatic blood at clinical sites. The purpose of the proposed method is to overcome the shortcomings of conventional methods and to accurately and quickly identify blood cells in mechanical images. The YOLOv7 model has been utilizedbecause it can detect objects in real-time with highprecision and speed. The methodology for bloodcell counting involves the use of YOLOv7 togenerate bounding boxes around each detectedblood cell in an image, where the count of these bounding boxes directly corresponds to the number of cells. This advancement in hematology not onlyimproves blood cell analysis efficiency but also expedites the process of diagnosing and planningtreatment in various medical situations. This paper details how YOLOv7 was modified and adjusted to meet the specific requirements of hematologicalimage analysis, including training models, preparing datasets, and evaluating performance. The study also addresses the potential impact of this technology on clinical workflows, highlightinghow it might help medical practitioners make decisions more quickly and intelligently. Hematology analysis systems including YOLOv7 are able to improve laboratory diagnostics and help patients by providing better care and results. This workconcludes by demonstrating the revolutionary potential of YOLOv7 in blood cell identificationand counting in hematology, paving the way for accurate and better

关键词： Analytical models Accuracy Computational modeling Cells (biology) Real-time systems Reliability Object recognition Medical diagnosis Blood Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

Accurate Anomaly Detection in Medical Images using Transfer Learning and Data Optimization: MRI and CT as Case Studies

Accurate Anomaly Detection in Medical Images using Transfer ...

引用

Neural Networks and Neurotechnologies (NeuroNT), II International Conference on

作者： Yasser Nizamli Anton Yu. Filatov Weaam Fadel Yulia A. Shichkina Department of Software Engineering and Computer Applications Saint Petersburg Electrotechnical University “LETI” Saint Petersburg Russian Federation Department of Computer Science Saint Petersburg Electrotechnical University “LETI” Saint Petersburg Russian Federation

ISBN: (数字)9798350363739

ISBN: (纸本)9798350363746

Today, medical imaging techniques are widely used to detect a variety of human conditions and diseases. To speed up the diagnostic process, systems are often automated using deep learning methods, which have been proven to yield outstanding results. However, these models may be limited to a specific domain, prone to overfitting, and difficult to update. In this paper, we present a transfer learning approach that achieves high accuracy for detecting anomalies in both MRI and CT images. Our system utilizes the ConvNeXt network, inspired by the Swin Transformer architecture, as a high-level feature extractor from processed and augmented medical images. The resulting data is then optimized using the Edited Nearest Neighbors algorithm, which selects the most effective observations and discards the rest. Finally, the data is passed to the K-Nearest Neighbors classifier to detect anomalies in numerically represented images. Our model achieves high performance for two different types of medical images and can easily be updated without the need for retraining.

关键词： Accuracy Magnetic resonance imaging Computed tomography Transfer learning Feature extraction Brain modeling Numerical models

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：