Procedia Computer Science最新文献

英文中文

Enhancing Sustainable Supply Chain Forecasting Using Machine Learning for Sales Prediction

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.006

Md. Parvezur Rahman Mahin , Munem Shahriar , Ritu Rani Das , Anuradha Roy , Ahmed Wasif Reza

Managing the supply chain is crucial to success in the competitive business sector. Demand forecasting using sales data is one of the major things in supply chain management because it is directly connected to profit margins, inventory levels, sales, and customer satisfaction. This research tried to provide an innovative approach to sales prediction using advanced machine learning methods to enhance supply chain operations and boost the predictive accuracy of supply chain models after analyzing historical sales data and considering different factors like seasonality, trends, and stock. Various machine learning algorithms were applied, including Linear Regression, Elastic Net Regression, KNN, Random Forest, and the ensemble Voting Regressor. The performance of Random Forest and KNN is very well but the Voting Regressor is better than other models for its strength of multiple algorithms. The Voting Regressor provides the lowest RMSE of 1.54 and the highest R² of 0.9999. This ensemble method improves sales forecasting accuracy by reducing errors and ensuring computational efficiency. It also provides more reliable tools to manage inventory, prevent overstocks, and minimize holding costs. This research presents the importance of machine learning integration in supply chain management. It shows the Voting Regressor as the most effective approach for demand forecast. Future research could explore the model’s application in broader markets, integrating other key factors and deep learning algorithms to refine predictive capabilities later.

{"title":"Enhancing Sustainable Supply Chain Forecasting Using Machine Learning for Sales Prediction","authors":"Md. Parvezur Rahman Mahin , Munem Shahriar , Ritu Rani Das , Anuradha Roy , Ahmed Wasif Reza","doi":"10.1016/j.procs.2025.01.006","DOIUrl":"10.1016/j.procs.2025.01.006","url":null,"abstract":"<div><div>Managing the supply chain is crucial to success in the competitive business sector. Demand forecasting using sales data is one of the major things in supply chain management because it is directly connected to profit margins, inventory levels, sales, and customer satisfaction. This research tried to provide an innovative approach to sales prediction using advanced machine learning methods to enhance supply chain operations and boost the predictive accuracy of supply chain models after analyzing historical sales data and considering different factors like seasonality, trends, and stock. Various machine learning algorithms were applied, including Linear Regression, Elastic Net Regression, KNN, Random Forest, and the ensemble Voting Regressor. The performance of Random Forest and KNN is very well but the Voting Regressor is better than other models for its strength of multiple algorithms. The Voting Regressor provides the lowest RMSE of 1.54 and the highest R<sup>2</sup> of 0.9999. This ensemble method improves sales forecasting accuracy by reducing errors and ensuring computational efficiency. It also provides more reliable tools to manage inventory, prevent overstocks, and minimize holding costs. This research presents the importance of machine learning integration in supply chain management. It shows the Voting Regressor as the most effective approach for demand forecast. Future research could explore the model’s application in broader markets, integrating other key factors and deep learning algorithms to refine predictive capabilities later.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 470-479"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377150","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Exploring Ransomware Detection Based on Artificial Intelligence and Machine Learning

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.014

Mayur Rele , John Samuel , Dipti Patil , Udaya Krishnan

Ransomware is an increasingly prevalent cybersecurity hazard due to its ability to encrypt data and request payment for its decryption. The threat’s dynamic nature generally renders conventional ransomware detection methods ineffective. This paper suggests an innovative method for detecting ransomware that capitalizes on artificial intelligence (AI) and machine learning (ML). A novel technique has been developed that integrates robust anomaly detection and classification algorithms with advanced feature extraction from system logs, network traffic, and file metadata. This technique achieves high accuracy with minimal false-positive rates by employing autoencoders, isolated forests for anomaly detection, random forests, and support vector machines for classification. The method’s ability to substantially improve ransomware defenses has been demonstrated through extensive testing on a large dataset, revealing that it outperforms current approaches. The study establishes a firm foundation for proactive ransomware detection and mitigation by demonstrating the advantages of integrating AI and ML in cybersecurity.

{"title":"Exploring Ransomware Detection Based on Artificial Intelligence and Machine Learning","authors":"Mayur Rele , John Samuel , Dipti Patil , Udaya Krishnan","doi":"10.1016/j.procs.2025.01.014","DOIUrl":"10.1016/j.procs.2025.01.014","url":null,"abstract":"<div><div>Ransomware is an increasingly prevalent cybersecurity hazard due to its ability to encrypt data and request payment for its decryption. The threat’s dynamic nature generally renders conventional ransomware detection methods ineffective. This paper suggests an innovative method for detecting ransomware that capitalizes on artificial intelligence (AI) and machine learning (ML). A novel technique has been developed that integrates robust anomaly detection and classification algorithms with advanced feature extraction from system logs, network traffic, and file metadata. This technique achieves high accuracy with minimal false-positive rates by employing autoencoders, isolated forests for anomaly detection, random forests, and support vector machines for classification. The method’s ability to substantially improve ransomware defenses has been demonstrated through extensive testing on a large dataset, revealing that it outperforms current approaches. The study establishes a firm foundation for proactive ransomware detection and mitigation by demonstrating the advantages of integrating AI and ML in cybersecurity.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 548-556"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376858","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Assessing the Significance of Machine Learning in Forecasting Energy Recovery from Waste

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.022

Nirzona Binta Badal , Fahmida Anjum , Ismail Mahmud Nur , Tithi Paul , Ahmed Wasif Reza

Accurate estimating energy recovery from trash is vital for optimizing the Waste-to-Energy (WTE) mechanisms essential in tackling global waste management processes and energy sustainability issues. This paper analyzes a synthetically expanded dataset generated with the help of SMOTE techniques to compare the performance of four machine learning (ML) models, Decision Tree Regression, Random Forest Regression, CatBoost Regression, and XGBoost Regression. Here, the dataset contains important waste parameters like composition, moisture content, and treatment procedures that help the models forecast energy output with high precision—datasets and evaluate additional machine learning techniques to boost prediction accuracy in industrial WTE systems further. Performance indicators like MAE, RMSE, MAPE, and R² scores have been assessed here to identify each model’s accuracy and computational efficiency. The final result of the analysis states that the ensemble based models, more precisely XG-Boost and CatBoost, outperformed the simpler ones like Decision Tree, where CatBoost achieved the best R² value of 0.9893 and the minimum MAPE of 12.90 percent. Though using little extra storage, CatBoost showed great performance. The obtained results bring useful insights into efficient model selection for WTE applications. Further studies shall therefore be exclusively focused on validating this study’s results in real life conditions.

{"title":"Assessing the Significance of Machine Learning in Forecasting Energy Recovery from Waste","authors":"Nirzona Binta Badal , Fahmida Anjum , Ismail Mahmud Nur , Tithi Paul , Ahmed Wasif Reza","doi":"10.1016/j.procs.2025.01.022","DOIUrl":"10.1016/j.procs.2025.01.022","url":null,"abstract":"<div><div>Accurate estimating energy recovery from trash is vital for optimizing the Waste-to-Energy (WTE) mechanisms essential in tackling global waste management processes and energy sustainability issues. This paper analyzes a synthetically expanded dataset generated with the help of SMOTE techniques to compare the performance of four machine learning (ML) models, Decision Tree Regression, Random Forest Regression, CatBoost Regression, and XGBoost Regression. Here, the dataset contains important waste parameters like composition, moisture content, and treatment procedures that help the models forecast energy output with high precision—datasets and evaluate additional machine learning techniques to boost prediction accuracy in industrial WTE systems further. Performance indicators like MAE, RMSE, MAPE, and R² scores have been assessed here to identify each model’s accuracy and computational efficiency. The final result of the analysis states that the ensemble based models, more precisely XG-Boost and CatBoost, outperformed the simpler ones like Decision Tree, where CatBoost achieved the best R² value of 0.9893 and the minimum MAPE of 12.90 percent. Though using little extra storage, CatBoost showed great performance. The obtained results bring useful insights into efficient model selection for WTE applications. Further studies shall therefore be exclusively focused on validating this study’s results in real life conditions.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 623-632"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376902","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Evaluation of Multiple Apache Spark Applications using Kubernetes as a Cluster manager on Google Cloud

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.017

M. Jayanthi Dr. , K. Ram Mohan Rao Dr. , Vuppala Sukanya

Big data processing frameworks demands for scalable and efficient cluster management. Apache Spark has emerged as prominent big data processing framework providing high-speed data processing and analytics capabilities for multiple applications. This paper explores the integration of Kubernetes as a cluster manager for Apache Spark applications leveraging its containerization capabilities to improve resource utilization and simplify deployment. In this paper the challenges of deploying spark applications on traditional cluster managers and showcase the advantages of adopting Kubernetes are analysed. The experimental evaluation demonstrates the benefits of Kubernetes as a cluster manager for Apache Spark framework. To execute the multiple Apache Spark applications on Kubernetes a homogenous cluster on Google Cloud is created by History bucket and service account. Finally multiple applications are executed on Google Kubernetes Engine. Output can be shown as the number of executor pods created with the performance metrics can be viewed. In conclusion, this paper compares the performance metrics such as job execution time and resource utilization with the different cluster.

{"title":"Evaluation of Multiple Apache Spark Applications using Kubernetes as a Cluster manager on Google Cloud","authors":"M. Jayanthi Dr. , K. Ram Mohan Rao Dr. , Vuppala Sukanya","doi":"10.1016/j.procs.2025.01.017","DOIUrl":"10.1016/j.procs.2025.01.017","url":null,"abstract":"<div><div>Big data processing frameworks demands for scalable and efficient cluster management. Apache Spark has emerged as prominent big data processing framework providing high-speed data processing and analytics capabilities for multiple applications. This paper explores the integration of Kubernetes as a cluster manager for Apache Spark applications leveraging its containerization capabilities to improve resource utilization and simplify deployment. In this paper the challenges of deploying spark applications on traditional cluster managers and showcase the advantages of adopting Kubernetes are analysed. The experimental evaluation demonstrates the benefits of Kubernetes as a cluster manager for Apache Spark framework. To execute the multiple Apache Spark applications on Kubernetes a homogenous cluster on Google Cloud is created by History bucket and service account. Finally multiple applications are executed on Google Kubernetes Engine. Output can be shown as the number of executor pods created with the performance metrics can be viewed. In conclusion, this paper compares the performance metrics such as job execution time and resource utilization with the different cluster.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 576-582"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376956","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Conditional GANs in Image-to-Image Translation: Improving Accuracy and Contextual Relevance in Diverse Datasets

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.056

Smit Gandhi , Hemit Rana , Nikita Bhatt

Image-to-image translation involves converting images from one domain to another while preserving key features, making it useful for tasks like style transfer, colorization, and super-resolution. By learning the underlying relationships between different domains, this method enables the creation of realistic images that conform to the desired style. This study investigates the application of Generative Adversarial Networks (GANs) to image-to-image translation. However, experiments show that GANs often struggle to fully capture the diversity of the data, which is crucial for accurate translation between image domains. To address this limitation, the research shifts focus to Conditional Generative Adversarial Networks (CGANs), exploring their potential for overcoming the shortcomings of traditional GANs. The findings demonstrate that CGANs outperform GANs in generating high-quality images across a range of data types. By conditioning the generation process on additional input data, CGANs improve both the quality and accuracy of the generated images. This conditional approach ensures better alignment between generated outputs and input labels, leading to more consistent and precise image translations. The observed improvements highlight CGANs as a more effective model for applications requiring detailed and accurate image generation.

{"title":"Conditional GANs in Image-to-Image Translation: Improving Accuracy and Contextual Relevance in Diverse Datasets","authors":"Smit Gandhi , Hemit Rana , Nikita Bhatt","doi":"10.1016/j.procs.2025.01.056","DOIUrl":"10.1016/j.procs.2025.01.056","url":null,"abstract":"<div><div>Image-to-image translation involves converting images from one domain to another while preserving key features, making it useful for tasks like style transfer, colorization, and super-resolution. By learning the underlying relationships between different domains, this method enables the creation of realistic images that conform to the desired style. This study investigates the application of Generative Adversarial Networks (GANs) to image-to-image translation. However, experiments show that GANs often struggle to fully capture the diversity of the data, which is crucial for accurate translation between image domains. To address this limitation, the research shifts focus to Conditional Generative Adversarial Networks (CGANs), exploring their potential for overcoming the shortcomings of traditional GANs. The findings demonstrate that CGANs outperform GANs in generating high-quality images across a range of data types. By conditioning the generation process on additional input data, CGANs improve both the quality and accuracy of the generated images. This conditional approach ensures better alignment between generated outputs and input labels, leading to more consistent and precise image translations. The observed improvements highlight CGANs as a more effective model for applications requiring detailed and accurate image generation.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 954-963"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Implementation of an Asset Tracking Embedded System using Ultra-Wide Band Technology

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2024.12.035

Gerardine Immaculate Mary , CH S K N S Dixit , Bapatla Venkata Naresh , Anitha Julian

In the most complex indoor environments, such as a warehouse, where there is a large volume of equipment and the possibility of losing it is inevitable, this study reveals an embedded system that uses Ultra-Wide Band (UWB) accuracy to deliver pinpoint asset tracking. It is crucial to keep an eye on them. constructing a Real-Time Locating System (RTLS), to use technical terminology. Even though they were effective, conventional navigation systems like GPS, Wi-Fi, and BLE have limitations due to interference and range. UWB effectively overcomes these constraints with its massive 7.5GHz bandwidth and lower interference (- 41.3dBm/1MHz). When it comes to interior navigation, UWB works better since it has a higher connection than Wi-Fi and GPS. Conversely, UWB has no propagation effects like as scattering or reflection and can simply flow through an obstruction. The goal is to create an embedded system that uses an Arduino UNO to gather data from a Deca Wave DW1000, which will serve as a tag and anchor for effective asset navigation utilizing the Time of Flight (ToF) localization method.

{"title":"Implementation of an Asset Tracking Embedded System using Ultra-Wide Band Technology","authors":"Gerardine Immaculate Mary , CH S K N S Dixit , Bapatla Venkata Naresh , Anitha Julian","doi":"10.1016/j.procs.2024.12.035","DOIUrl":"10.1016/j.procs.2024.12.035","url":null,"abstract":"<div><div>In the most complex indoor environments, such as a warehouse, where there is a large volume of equipment and the possibility of losing it is inevitable, this study reveals an embedded system that uses Ultra-Wide Band (UWB) accuracy to deliver pinpoint asset tracking. It is crucial to keep an eye on them. constructing a Real-Time Locating System (RTLS), to use technical terminology. Even though they were effective, conventional navigation systems like GPS, Wi-Fi, and BLE have limitations due to interference and range. UWB effectively overcomes these constraints with its massive 7.5GHz bandwidth and lower interference (- 41.3dBm/1MHz). When it comes to interior navigation, UWB works better since it has a higher connection than Wi-Fi and GPS. Conversely, UWB has no propagation effects like as scattering or reflection and can simply flow through an obstruction. The goal is to create an embedded system that uses an Arduino UNO to gather data from a Deca Wave DW1000, which will serve as a tag and anchor for effective asset navigation utilizing the Time of Flight (ToF) localization method.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 331-340"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376668","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Improving Breast Cancer Diagnosis through Advanced Image Analysis and Neural Network Classifications

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2024.12.008

Kanagamalliga S , Dandu Bhavya Varma

Breast cancer, one of the deadly cancers affecting women globally, is caused by genetic mutations that affect cell growth and division. Limitations are present in traditional recognition methods such as mammography, image fusion, and convolutional neural networks (CNNs) in accurately distinguishing between benign, malignant, and normal breast tissues. An enhanced recognition system utilizing advanced image analysis techniques and neural network classifications is proposed by this research to improve diagnostic accuracy. A combination of grey level co-occurrence matrix (GLCM) for texture feature extraction, Expectation Maximization based Gaussian Mixture Model (EMGMM) segmentation, and K-means clustering algorithms are employed by the proposed system. Robust analysis and classification of breast tissue images are provided by the integration of these methods, offering a more precise differentiation between benign, malignant, and normal conditions. The reduction of errors associated with traditional screening methods, enhanced noise recognition, and a non-invasive approach compared to conventional biopsy techniques are included among the key benefits of this system. Improved precision and reliability in breast cancer diagnosis are achieved through the use of neural networks combined with advanced image analysis algorithms. The time and discomfort associated with traditional diagnostic procedures are also reduced by this system, making it a more user-friendly option for patients. The early recognition and treatment of breast cancer are aimed to be enhanced by leveraging these advanced techniques, contributing to lower mortality rates. The potential of integrating GLCM, EMGMM segmentation, and K-means clustering with neural networks, providing a more effective solution for breast cancer screening and diagnosis. Significant improvements in the efficiency of breast cancer diagnostics are promised by this innovative approach, achieving 96.8% accuracy.

{"title":"Improving Breast Cancer Diagnosis through Advanced Image Analysis and Neural Network Classifications","authors":"Kanagamalliga S , Dandu Bhavya Varma","doi":"10.1016/j.procs.2024.12.008","DOIUrl":"10.1016/j.procs.2024.12.008","url":null,"abstract":"<div><div>Breast cancer, one of the deadly cancers affecting women globally, is caused by genetic mutations that affect cell growth and division. Limitations are present in traditional recognition methods such as mammography, image fusion, and convolutional neural networks (CNNs) in accurately distinguishing between benign, malignant, and normal breast tissues. An enhanced recognition system utilizing advanced image analysis techniques and neural network classifications is proposed by this research to improve diagnostic accuracy. A combination of grey level co-occurrence matrix (GLCM) for texture feature extraction, Expectation Maximization based Gaussian Mixture Model (EMGMM) segmentation, and K-means clustering algorithms are employed by the proposed system. Robust analysis and classification of breast tissue images are provided by the integration of these methods, offering a more precise differentiation between benign, malignant, and normal conditions. The reduction of errors associated with traditional screening methods, enhanced noise recognition, and a non-invasive approach compared to conventional biopsy techniques are included among the key benefits of this system. Improved precision and reliability in breast cancer diagnosis are achieved through the use of neural networks combined with advanced image analysis algorithms. The time and discomfort associated with traditional diagnostic procedures are also reduced by this system, making it a more user-friendly option for patients. The early recognition and treatment of breast cancer are aimed to be enhanced by leveraging these advanced techniques, contributing to lower mortality rates. The potential of integrating GLCM, EMGMM segmentation, and K-means clustering with neural networks, providing a more effective solution for breast cancer screening and diagnosis. Significant improvements in the efficiency of breast cancer diagnostics are promised by this innovative approach, achieving 96.8% accuracy.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 73-80"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376725","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning approach for weather classification using pre-trained convolutional neural networks

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2024.12.015

Harit Tarwani , Shivang Patel , Parth Goel

Machine Learning has had significant uses in weather prediction, and is especially useful in areas of agriculture and livestock, and outdoor vision. While various studies have been done on weather forecasting, limited work has yet been undertaken on its classification. Automating weather classification can free and save human resources. While simultaneously enhancing decision-making in the fields discussed. As such, the researchers have implemented transfer learning on different convolutional neural network (CNN) models, like ResNet50V2, EfficientNetB5, and EfficientNetV2S, for classification across 11 fields. The models were fine-tuned to improve performance and trained on a dataset that included diverse conditions like snow, lightning, and haze. Among the considered models, EfficientNetV2S showed the best accuracy (92.35%). This outcome indicates the effective application of machine learning, mainly transfer learning, in the future for weather classification and thus, its subsequent uses.

引用次数: 0

Integrating NAS for Human Pose Estimation

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2024.12.020

Srilakshmi V, Uday Kiran G, B Moulika, G S Mahitha, G Laukya, M Ruthick

Neural Architecture Search (NAS) technologies have become popular across various fields, allowing for the joint learning of neural network architectures and weights. However, most existing NAS methods are task-specific, focusing on optimizing a single architecture to replace human-designed networks while often neglecting domain knowledge. This paper introduces Pose Neural Fabrics Search (PoseNFS), a unique NAS framework that integrates domain knowledge via part-specific neural architecture search—a form of multi-task learning—for human posture estimation.

PoseNFS utilizes a novel search space called Cell-based Neural Fabric (CNF), employing a differentiable search approach to facilitate learning at both micro and macro levels. By utilizing prior knowledge of human body structure, PoseNFS directs the search for part-specific architectures personalized to different body components, treating the localization of human key points as multiple disentangled sub-tasks. Experimental results on the MPII and MS-COCO datasets demonstrate that PoseNFS significantly outperforms a manually designed part-based baseline model and several state-of-the-art methods, validating the effectiveness of this knowledge-guided strategy.

{"title":"Integrating NAS for Human Pose Estimation","authors":"Srilakshmi V, Uday Kiran G, B Moulika, G S Mahitha, G Laukya, M Ruthick","doi":"10.1016/j.procs.2024.12.020","DOIUrl":"10.1016/j.procs.2024.12.020","url":null,"abstract":"<div><div>Neural Architecture Search (NAS) technologies have become popular across various fields, allowing for the joint learning of neural network architectures and weights. However, most existing NAS methods are task-specific, focusing on optimizing a single architecture to replace human-designed networks while often neglecting domain knowledge. This paper introduces Pose Neural Fabrics Search (PoseNFS), a unique NAS framework that integrates domain knowledge via part-specific neural architecture search—a form of multi-task learning—for human posture estimation.</div><div>PoseNFS utilizes a novel search space called Cell-based Neural Fabric (CNF), employing a differentiable search approach to facilitate learning at both micro and macro levels. By utilizing prior knowledge of human body structure, PoseNFS directs the search for part-specific architectures personalized to different body components, treating the localization of human key points as multiple disentangled sub-tasks. Experimental results on the MPII and MS-COCO datasets demonstrate that PoseNFS significantly outperforms a manually designed part-based baseline model and several state-of-the-art methods, validating the effectiveness of this knowledge-guided strategy.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 182-191"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376754","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Application of Hyperledger Blockchain Technology to Logistics Supply Chain with IoT

Procedia Computer Science

Pub Date : 2025-01-01 DOI: 10.1016/j.procs.2025.01.042

Ahmet Sayar , Muhammet Cüneyd Kurtbaş , Çağlayan Sancaktar , Rana Dudu Kabak , Şükrü Çakmak

Product movement tracking stands as a pivotal aspect of logistics operations. The ever-expanding application of IoT technologies holds significant potential to enhance product tracking in logistics. However, the rapid expansion of IoT usage has also introduced several challenges, particularly concerning data privacy, security, and transparency. Blockchain technology, through distributed ledger systems, offers a secure and transparent means to manage transactions within a network. In this study, the Hyperledger Fabric blockchain network, chosen for its private and permissioned nature, leverages authentication features to ensure the confidentiality, integrity, and security of data from IoT devices. By integrating the widely adopted MQTT protocol, IoT devices involved in logistics processes communicate effectively with the blockchain network. This study presents a system where IoT devices enable the secure and automated tracking of data through blockchain’s registration and distribution mechanisms, all of which are demonstrated within a simulation environment. Developed specifically for the logistics sector, this system holds promising potential to address challenges across various fields.

{"title":"Application of Hyperledger Blockchain Technology to Logistics Supply Chain with IoT","authors":"Ahmet Sayar , Muhammet Cüneyd Kurtbaş , Çağlayan Sancaktar , Rana Dudu Kabak , Şükrü Çakmak","doi":"10.1016/j.procs.2025.01.042","DOIUrl":"10.1016/j.procs.2025.01.042","url":null,"abstract":"<div><div>Product movement tracking stands as a pivotal aspect of logistics operations. The ever-expanding application of IoT technologies holds significant potential to enhance product tracking in logistics. However, the rapid expansion of IoT usage has also introduced several challenges, particularly concerning data privacy, security, and transparency. Blockchain technology, through distributed ledger systems, offers a secure and transparent means to manage transactions within a network. In this study, the Hyperledger Fabric blockchain network, chosen for its private and permissioned nature, leverages authentication features to ensure the confidentiality, integrity, and security of data from IoT devices. By integrating the widely adopted MQTT protocol, IoT devices involved in logistics processes communicate effectively with the blockchain network. This study presents a system where IoT devices enable the secure and automated tracking of data through blockchain’s registration and distribution mechanisms, all of which are demonstrated within a simulation environment. Developed specifically for the logistics sector, this system holds promising potential to address challenges across various fields.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 814-823"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

首页上一页

下一页尾页

类型

全部化学•材料生命科学医学物理工程技术环境•农林材料科学地球科学法学管理学化学环境科学与生态学计算机科学教育学经济学农林科学人文科学生物学数学物理与天体物理心理学综合性期刊其他工业工程理学历史学农学文学信息工程

数据库

全部 ACS Publications Elsevier ieeexplore Springer The Royal Society of Chemistry Wiley

期刊

Procedia Computer Science

全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.

﹀