检索结果-内蒙古大学图书馆

4th International Conference on Intelligent Technologies, CONIT 2024

作者： Priya, M.Yagnasri Shendre, Sanjivani P. Pati, Peeta Basa Amrita School of Computing Department of Computer Science and Engineering Amrita Vishwa Vidyapeetham Bengaluru India

ISBN: (纸本)9798350349900

Gunshot detection and classification are pivotal for ensuring public safety, law enforcement, defense, and forensic investigations. This paper presents a comparitive study into the effectiveness of deep learning and traditional machine learning models in classifying gunshot sounds through Short-Term Fourier Transform (STFT) spectograms is presented. The study involves integrating Convolutional Neural Networks (CNNs) with standard machine learning techniques to extract and analyze features, particularly focusing on the penultimate layer of CNNs. By leveraging these features, this paper aims to enhance the system's ability to accurately identify complex sound patterns associated with gunshots. This analysis provides insights into the strengths and weaknesses of deep learning and machine learning approaches in gunshot classification, particularly achieving a perfect accuracy score with AlexNet and XGBoost. These findings pave the way for more effective gunshot detection systems in a variety of operational scenarios, thereby significantly contributing to public safety measures. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Flex Sensor Data Analysis for Hand Rehabilitation using Wearable Glove 3

Flex Sensor Data Analysis for Hand Rehabilitation using Wear...

引用

3rd International Conference on Electrical, Electronics, Information and Communication Technologies, ICEEICT 2024

作者： Thangam, S. Reddy, Anusha C. Varun, Ginnaram Hemasri, M. Virupaxappa Kumari, J. Jesy Janet Amrita Vishwa Vidyapeetham Amrita School of Computing Dept of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350369083

Every day we encounter many people with unexpected situations where they suffer from many neuropathological diseases that cause a severe impact on the daily lives of individuals, especially about hand movements, thereby directly affecting the quality of their lives. Traditional methods often lack insights into personalized progress and optimal medication outcomes, which take longer to recover. We have designed an approach based on a novel concept using wearable technology to monitor and assess finger movements. The proposed solution of wearable device integrated flex sensor technology to enhance the rehabilitation process by providing real-time insights by the input of the flex sensors captured by the finger movements of the patients. By sensor data analysis, people get a chance to know their progress and the professionals gain accurate tracking of the progress so that they enhance identifying the areas of improvement and customize the treatment. This innovative approach to designing a wearable device serves as a helping aid to improve patient outcomes by categorizing the stages of improvement into Normal, Mild, and Severe Impairment for a better understanding and helps improve the rehabilitation experience. © 2024 IEEE.

关键词： Personalized medicine

来源：评论

学校读者我要写书评

暂无评论

Enhancing Software Test Effort Estimation using Ensemble Learning Algorithms 4

Enhancing Software Test Effort Estimation using Ensemble Lea...

引用

4th International Conference on Intelligent Technologies, CONIT 2024

作者： Venkatraman, K. Akashvarma, M. Siddharth, S. Amrita School of Computing Department of Computer Science & Engineering Amrita Vishwa Vidyapeetham Chennai India

ISBN: (纸本)9798350349900

In software engineering, accurate test effort prediction is critical to project schedule and resource efficiency. Conventional approaches, which depend on past performance or expert opinion, frequently fail to provide a reliable estimate of the testing effort. These approaches might offer some insight at first, but as project complexity rises, they become less scalable and more error-prone. This study explores the use of machine learning to improve the estimation of software test effort in response. This work intends to overcome the shortcomings of conventional methods and create more accurate and scalable solutions by utilising the power of machine learning algorithms. The goal of the research is to determine the best machine learning strategy for software test effort estimation through extensive testing and comparative analysis. The results of this study are compared using metrics such as RMSE, MSE, MAE, and R2 score, with Random Forest algorithm demonstrating superior performance over other algorithms. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Optimizing Smart Contract Security: A Cost-Sensitive Graph Neural Network Approach for Vulnerability Detection

Optimizing Smart Contract Security: A Cost-Sensitive Graph N...

引用

2024 IEEE International Conference on Electronic Systems and Intelligent computing, ICESIC 2024

作者： Aravind, M. Saravanan, R. Saranya, G. Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science Engineering Chennai India

ISBN: (数字)9798331522988

ISBN: (纸本)9798331522988

This paper addresses the issues of blockchain platforms, especially Ethereum, smart contract vulnerabilities have resulted in significant monetary losses. The primary objective of this research is to identify critical smart contract vulnerabilities, such as reentrancy, overflow, and access control problems, by utilizing Graph Neural Networks (GNNs). To evaluate the degree to which two GNN-based models work at reducing practical financial losses, two of them are created using various training approaches. The first model treats each vulnerability class equally in an effort to maximize accuracy. The second model takes a more cost-effective tack by giving training greater weight to major weaknesses. We investigate the trade-offs between accuracy and financial risk reduction using a dataset of 2,217 contracts with imbalanced class distribution, replicating real-world settings. Financial loss estimates are derived from industry-reported incidents, and confusion matrices are employed to demonstrate the detection patterns. The results show that even though the first model's overall accuracy is higher at 96%, it endures more financial loss as a result of missing vital vulnerabilities. With a 94% accuracy rate, on the other hand, the cost-sensitive model effectively lowers high-impact false negatives, resulting in a less overall financial loss. Given that even a slight drop in accuracy might lead to more successful financial risk mitigation, our findings emphasize the significance of prioritizing vulnerability identification based on possible real-world repercussions. The groundwork for incorporating cost-sensitive models into smart contract auditing tools to improve blockchain security is provided by this research. © 2024 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Machine Learning Models for Water Quality Prediction 4

Comparative Analysis of Machine Learning Models for Water Qu...

引用

4th International Conference on Advances in Electrical, computing, Communication and Sustainable Technologies, ICAECT 2024

作者： Krushna, Bode Vamsi Sasikala, D. Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350343670

Water Quality plays an essential role in assessing environmental health, which has profound implications for both ecosystems and human well-being. In this work, various machine learning models including Logistic Regression, Decision Trees, Gaussian Naive Bayes, Random Forest, Support Vector Machine (SVM), and XGBoost were employed to analyze and predict water quality. The dataset used for the analysis encompasses diverse water quality parameters such as hardness, pH levels, solids, chloramines, turbidity, sulfate among others. Class imbalances in the dataset is addressed with Synthetic Minority Over-sampling Technique(SMOTE). The research findings highlight the significant improvement in classification accuracy of underrepresented data points due to SMOTE, with SVM and XGBoost achieving a training accuracy of around 78% and 75%, respectively. The study used K-fold cross-validation techniques and strategies to handle convergence warnings, enhancing the credibility and adaptability of the predictions. This study not only contributes to the growing body of knowledge in water quality analysis but also provides practical insights for researchers and practitioners seeking to employ machine learning approaches in environmental monitoring, water contamination. These findings highlight the critical role of machine learning in facilitating informed decision-making for long-term water management, ensuring the health and longevity of communities and ecosystems. © 2024 IEEE.

关键词： Decision Trees Gaussian Naive Bayes Logistic Regression Machine learning Random Forest SVM Water quality management XGBoost

来源：评论

学校读者我要写书评

暂无评论

Ransome Ware Detection Using Machine Learning and Deep Learning Models 1

引用

12th International Conference on Recent Trends in computing, ICRTC 2024

作者： Vivek, Vinnakota Sai Lokesh, Vadlamudi Kaushik, Tata Kumaran, U. Umah, B. Devarajan, Gopal Ganesh Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India Department of Computer Science and Engineering SRM Institute of Science and Technology Delhi NCR Campus Uttar Pradesh Ghaziabad India

ISBN: (数字)9789819789467

ISBN: (纸本)9789819789450

This research addresses the critical challenge of ransomware detection through the use of deep learning and machine learning methods. Because ransomware is a serious threat to cybersecurity, it is imperative that advanced techniques be used for prompt detection and mitigation. Using a large dataset, the study concentrates on pertinent features necessary for efficient machine learning model training. The suggested model shows promising results in identifying and averting ransomware attacks by utilizing cutting-edge algorithms. The model’s effectiveness is confirmed by the evaluation metrics (F1 score, recall, accuracy, and precision). By providing a proactive defense against emerging ransomware threats, the findings enhance cybersecurity measures. The study not only provides a strong methodology for ransomware detection but also encourages more research in the area with the goal of continuously bolstering cyber defenses against evolving cyber threats. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Learning in sinusoidal spaces with physics-informed neural networks

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第3期5卷 985-1000页

作者： Wong, Jian Cheng Ooi, Chin Chun Gupta, Abhishek Ong, Yew Soon Institute of High Performance Computing Agency for Science Technology and Research Singapore138632 Singapore School of Computer Science and Engineering Nanyang Technological University Singapore639798 Singapore Singapore Institute of Manufacturing Technology Agency for Science Technology and Research Singapore138632 Singapore Agency for Science Technology and Research Singapore138632 Singapore Data Science and Artificial Intelligence Research Centre School of Computer Science and Engineering Nanyang Technological University Singapore138632 Singapore

A physics-informed neural network (PINN) uses physics-Augmented loss functions, e.g., incorporating the residual term from governing partial differential equations (PDEs), to ensure its output is consistent with fundamental physics laws. However, it turns out to be difficult to train an accurate PINN model for many problems in practice. In this article, we present a novel perspective of the merits of learning in sinusoidal spaces with PINNs. By analyzing behavior at model initialization, we first show that a PINN of increasing expressiveness induces an initial bias around flat output functions. Notably, this initial solution can be very close to satisfying many physics PDEs, i.e., falling into a localminimum of the PINN loss that onlyminimizes PDE residuals, while still being far from the true solution that jointly minimizes PDE residuals and the initial and/or boundary conditions. It is difficult for gradient descent optimization to escape from such a local minimum trap, often causing the training to stall. We then prove that the sinusoidalmapping of inputs-in an architecture we label as sf-PINN-is effective to increase input gradient variability, thus avoiding being trapped in such deceptive local minimum. The level of variability can be effectively modulated to match high-frequency patterns in the problem at hand. A key facet of this article is the comprehensive empirical study that demonstrates the efficacy of learning in sinusoidal spaces with PINNs for a wide range of forward and inversemodeling problems spanning multiple physics domains. Impact Statement-Falling under the emerging field of physicsinformed machine learning, PINN models have tremendous potential as a unifying AI framework for assimilating physics theory and measurement data. However, they remain infeasible for broad science and engineering applications due to computational cost and training challenges, especially for more complex problems. Instead of focusing on empirical demonstration of appli

关键词： Gradient methods

来源：评论

学校读者我要写书评

暂无评论

A Smart Pill Container for Improved Medication 5

A Smart Pill Container for Improved Medication

引用

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

作者： Thangam, S. Reddy, T. Sudeep Dhanush Krithin, T. Kumari, J. Jesy Janet Amrita School of Computing Amrita Vishwa Vidyapeetham Dept of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350379945

Medication non-adherence is a significant problem in healthcare, leading to the gradual worsening of patient health, increased hospitalizations, and higher costs. Current patient-driven solutions rely heavily on the reliability of patient compliance and self-reporting, which are often inaccurate, leaving real-time monitoring and intervention gaps. The proposed system of a smart pill box, integrated with a Raspberry Pi 4, weight sensors, and a camera, addresses these weaknesses. The weight sensors will detect medication intake and verify dosages by sensing changes in weight, while the camera will capture images of the pills. Using a self-made dataset with over 400 images of each of 10 different tablets, the system will employ deep learning to identify the pills and verify dosages, alerting patients and caregivers to anomalies such as skipped doses. This automated adherence monitoring system reduces reliance on self-reporting and increases tracking accuracy. The expected outcomes of this research include improved medication adherence, reduced medication errors, and enhanced monitoring mechanisms, ultimately leading to better healthcare provision and outcomes for patients. Innovating health technology in this manner promises significant improvements in patient care. © 2024 IEEE.

关键词： Drug dosage

来源：评论

学校读者我要写书评

暂无评论

Accent-Enhanced Automatic Speech Recognition for Multilingual Classrooms 15

Accent-Enhanced Automatic Speech Recognition for Multilingua...

引用

15th International Conference on computing Communication and Networking Technologies, ICCCNT 2024

作者： Jagadeeshvaran, M. Sarah Michelle, H. Shruti Bala, S. Srinivasa Vinayak, M. Deepika, T. Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9798350370249

In an increasingly interconnected world, universities face the challenge of hosting diverse students from different backgrounds. The biggest obstacle to effective communication and learning in multicultural classrooms is teacher and student diversity. To overcome this challenge, we propose the development of noise-enhanced automatic speech recognition (ASR), specifically for multilingual classrooms. The system is designed to provide instant feedback on lessons while preserving the subtleties of speech, thus improving understanding and creating a learning environment. Using the advanced capabilities of the Transformer model together with the convolutional neural network (CNN) highlight classifier, the proposed system provides powerful and accurate transcription capabilities. This article provides an overview of the design, implementation, and evaluation of voice-enhanced ASR, demonstrating its potential to transform communication and learning in a variety of educational settings. © 2024 IEEE.

关键词： Accent Enhancement Accent Nullification Automatic Speech Recognition Inclusive Learning Multilingual Classrooms Speech Transcription

来源：评论

学校读者我要写书评

暂无评论

Optimization Techniques for the Traveling Salesman Problem: A Study of Conglomerate of Algorithmic Approaches 1

Optimization Techniques for the Traveling Salesman Problem: ...

引用

1st International Conference for Women in computing, InCoWoCo 2024

作者： Niharika, Yamini Akhila, Thirumalaraju Vaishnavi, Vemula Sahay, Apurvanand Amrita School of Computing Amrita Vishwa Vidyapeetham Department of Computer Science and Engineering Bangalore India

ISBN: (纸本)9798331518943

This project investigates Python to study the Traveling Salesman Problem (TSP) and looks at five different algorithms that can be implemented: Brute Force, Greedy, Genetic, Dynamic Programming, and Divide and Conquer. Locating the shortest path that makes accurate stops in each city before returning to the origin city is the task assigned to the TSP. In order to identify the best solution, the Brute Force algorithm thoroughly looks through every combination of city sequences. The nearest unexplored city is chosen at each stage via the Greedy Algorithm, in contrast, which makes locally optimal decisions. By applying genetic operators like crossover and mutation, genetic algorithms generate new generations of possible paths by iteratively improving existing ones, mimicking the principles of natural selection. To cut down on unnecessary computations, dynamic programming divides the problem into smaller subproblems and stores the answers. © 2024 IEEE.

关键词： Genetic algorithms

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：