检索结果-内蒙古大学图书馆

EDSA-Ensemble: an Event Detection Sentiment Analysis Ensemble Architecture

IEEE Transactions on Affective Computing 2024年第02期16卷 1-18页

作者： PETRESCU, Alexandru TRUICA, Ciprian-Octavian APOSTOL, Elena-Simona PASCHKE, Adrian Computer Science and Engineering Department Faculty of Automatic Control and Computers National University of Science and Technology Politehnica Bucharest Bucharest Romania Fraunhofer Institute for Open Communication Systems Berlin Germany

As global digitization continues to grow, technology becomes more affordable and easier to use, and social media platforms thrive, becoming the new means of spreading information and news. Communities are built around sharing and discussing current events. Within these communities, users are enabled to share their opinions about each event. Using Sentiment Analysis to understand the polarity of each message belonging to an event, as well as the entire event, can help to better understand the general and individual feelings of significant trends and the dynamics on online social networks. In this context, we propose a new ensemble architecture, EDSAEnsemble (Event Detection Sentiment Analysis Ensemble), that uses Event Detection and Sentiment Analysis to improve the detection of the polarity for current events from Social Media. For Event Detection, we use techniques based on Information Diffusion taking into account both the time span and the topics. To detect the polarity of each event, we preprocess the text and employ several Machine and Deep Learning models to create an ensemble model. The preprocessing step includes several word representation models: raw frequency, TFIDF, Word2Vec, and Transformers. The proposed EDSA-Ensemble architecture improves the event sentiment classification over the individual Machine and Deep Learning models. Authors

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Pneumonia Detection from X-ray Images Using Convolutional Neural Networks 24

Pneumonia Detection from X-ray Images Using Convolutional Ne...

引用

24th International Conference on control Systems and computer science, CSCS 2023

作者： Dediu, Viorel Udrea, Andreea Faculty of Automatic Control and Computer Science Univesity Politehnica of Bucharest Department of Automatic Control and Systems Engineering Bucharest Romania

ISBN: (纸本)9798350313390

Pneumonia is one of the top causes of death in Romania and early detection of this disease improves the recovery chances and shortens the length of hospitalization. In this work, we develop a solution for automatic pneumonia detection based on convolutional neural networks. Four network models are investigated. They are trained on 4.163 images from a public dataset and tested on 530 images. The best results are obtained by one of the proposed models conducting to a sensitivity of 98.72%, an accuracy of 89.81%, and ROC 93.46%. Thus, this research proposes a lightweight screening tool that can help triaging the patients with pneumonia. © 2023 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Optimizing Artificial Neural Networks Trough Weight Adjustments 28th

Optimizing Artificial Neural Networks Trough Weight Adjustme...

引用

28th International Conference on Knowledge Based and Intelligent information and Engineering Systems, KES 2024

作者： Akber, Syed Muhammad Abrar Szczesna, Agnieszka Kazmi, Sadia Nishat Department of Computer Graphics Vision and Digital Systems Faculty of Automatic Control Electronics and Computer Science Silesian University of Technology Gliwice44-100 Poland Faculty of Automatic Control Electronics and Computer Science Silesian University of Technology Gliwice44-100 Poland

Longer training times pose a significant challenge in artificial neural networks (ANNs) as it may leads to increasing the computational costs and decreasing the effectiveness of the model. Therefore, it is imperative to reduce training times in ANNs to enhance the computational efficiency. The initialization of the weights between the layers in ANN plays a vital role in reducing training times. Appropriate weight initialization can help the network converge faster during the training by providing an optimum starting point for the network. Therefore, weight initialization techniques are essential for efficient training of ANNs. This paper revisits and implements different popular weight initialization techniques in ANNs and analyzes their impact on training time. Specifically, this paper implements Gaussian-based, Kaming-based, and Xavier-based weight initiation atop a popular DNN-based network. The experiments are conducted by employing a well-known dataset. The results show that the scenario when no weight initiation is applied consumed the highest training time, whereas different weight initiation techniques contribute in reducing the training times for the network. © 2024 The Authors.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

EEG Multi-Objective Feature Selection using a Genetic Procedure with Hybrid Mutation Operator 24

EEG Multi-Objective Feature Selection using a Genetic Proced...

引用

24th International Conference on control Systems and computer science, CSCS 2023

作者： Cimpanu, Corina Gheorge Asachi Technical University of Iai Faculty of Automatic Control and Computer Science Computer Science Department Iai Romania

ISBN: (纸本)9798350313390

computer science developments in medical data processing increase modeling accuracy, speed up data analysis, and obtain automated diagnostics. As a result, the role of machine learning algorithms becomes significant, as well as the need to implement faster and more accurate systems. For example, classification is a common component in computer-aided diagnostics, and choosing the proper set of predictors is critical for training a classifier model. Usually, medical datasets are highly dimensional, and their characteristical data correlations require reducing the set of features used for classification. Besides supervised grouping algorithms, genetic optimization procedures are well known for their robust verification of competing attributes by directly estimating their usefulness on a specific classifier. This paper discusses the assessment of Electroencephalogram data samples acquired during working memory load n-back scenarios using a new embedded feature selection procedure based on Multi-Objective Optimization via Genetic Algorithms enhanced with a hybrid mutation operator. Experimental results focus on both the hybrid mutation operator performances, and obtaining overall higher Support Vector Machine classification accuracy while reducing the number of selected features. © 2023 IEEE.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Real-Time Multi-object Tracking Using YOLOv8 and SORT on a SoC FPGA 21st

Real-Time Multi-object Tracking Using YOLOv8 and SORT on a...

引用

21st International Symposium on Applied Reconfigurable Computing, ARC 2025

作者： Danilowicz, Michal Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Krakow Poland

ISBN: (纸本)9783031879944

Multi-object tracking (MOT) is one of the most important problems in computer vision and a key component of any vision-based perception system used in advanced autonomous mobile robotics. Therefore, its implementation on low-power and real-time embedded platforms is highly desirable. Modern MOT algorithms should be able to track objects of a given class (e.g. people or vehicles). In addition, the number of objects to be tracked is not known in advance, and they may appear and disappear at any time, as well as be obscured. For these reasons, the most popular and successful approaches have recently been based on the tracking paradigm. Therefore, the presence of a high quality object detector is essential, which in practice accounts for the vast majority of the computational and memory complexity of the whole MOT system. In this paper, we propose an FPGA (Field-Programmable Gate Array) implementation of an embedded MOT system based on a quantized YOLOv8 detector and the SORT (Simple Online Realtime Tracker) tracker. We use a modified version of the FINN framework to utilize external memory for model parameters and to support operations necessary required by YOLOv8. We discuss the evaluation of detection and tracking performance using the COCO and MOT15 datasets, where we achieve 0.21 mAP and 38.9 MOTA respectively. As the computational platform, we use an MPSoC system (Zynq UltraScale+ device from AMD/Xilinx) where the detector is deployed in reprogrammable logic and the tracking algorithm is implemented in the processor system. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： System-on-chip

来源：评论

学校读者我要写书评

暂无评论

Ant Colony Optimization Algorithm for Finding the Maximum Number of d-Size Cliques in a Graph with Not All m Edges between Its d Parts 18th

Ant Colony Optimization Algorithm for Finding the Maximum Nu...

引用

Dependable computer Systems and Networks - Proceedings of the Eighteenth International Conference on Dependability of computer Systems DepCoS-RELCOMEX

作者： Schiff, Krzysztof Department of Automatic Control and Computer Science Cracow University of Technology Cracow31-155 Poland

ISBN: (纸本)9783031377198

A graph with m vertices in each of its d sections is a d-partite graph. In sections of a d-partite graph, the vertices are not connected by any edges. Each pair of vertices from different sections is always connected by an edge. If in a d-partite graph each section has m vertices, then such a graph can be divided into m d-vertex cliques if it has all edges between each pair of vertices from different sections. If in a graph with m vertices in each section there are no edges between each pair of vertices from different sections of the graph, then such a graph cannot be divided into m cliques. Such a graph can be easily divided into cliques, but their number does not have to be the maximum number of cliques into which such a graph can be divided. So this article deals with the problem of how to divide such a graph into cliques so that their number is the maximum number of cliques. For this purpose ant algorithms without and with two different dynamic desire function were used. The achieved results were compared and discussed. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Ant colony optimization

来源：评论

学校读者我要写书评

暂无评论

Partitioning of an M-Part Weighted Graph with N Vertices in Each Part into N Cliques with M Vertices and the Total Minimum Sum of Their Edges Weights Using Ant Algorithms 18th

Partitioning of an M-Part Weighted Graph with N Vertices in ...

引用

Dependable computer Systems and Networks - Proceedings of the Eighteenth International Conference on Dependability of computer Systems DepCoS-RELCOMEX

作者： Schiff, Krzysztof Department of Automatic Control and Computer Science Cracow University of Technology Cracow31-155 Poland

ISBN: (纸本)9783031377198

An m-part weighted graph with n vertices in each of its parts is used for tracking objects in multi-camera vision systems. This problem is called: The General Maximum Minimal Clique Problem. Different ant colony optimisation algorithms are presented to solve this problem. New dynamic desire functions are proposed and the results of their influence on the obtained total sum of the weight of cliques are presented. Thus the best dynamic desire functions were determined, which were determined to be better than the existing desire functions in terms of the obtained minimum sum of weights of all cliques. This problem occurs in the area of object tracking on film frames in cameras. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Cameras

来源：评论

学校读者我要写书评

暂无评论

Indirect Method for Calibrating Quadrotor Sensors: A Case Study applied to the Crazyflie 2.X 1

Indirect Method for Calibrating Quadrotor Sensors: A Case St...

引用

1st IEEE IAS Global Conference on Emerging Technologies, GlobConET 2022

作者： Socas, Rafael Dormido, Raquel Guinaldo, Maria Manas-Alvarez, Francisco Jose Dormido, Sebastian Department of Computer Science and Automatic Control Madrid Spain

ISBN: (数字)9781665443579

ISBN: (纸本)9781665443579

In this work, a novel algorithm for sensors calibration in quadrotor is presented. Classical procedures are used to calibrate some variables but for the others, the angular speeds, an indirect method has been proposed. The algorithm has been applied to the quadrotor Crazyflie 2.X in a practical way. The proposed method has been described in detail and it can be applied in other kind of quadrotors in a simple way. Some experiments in a real environment have been performed to validate the solution. This work is part of a framework for quadrotor control engineering that improves the global performance of the system. © 2022 IEEE.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Based Calibration Approaches for Single-Beam and Multiple-Beam Distance Sensors

引用

IEEE Sensors Journal 2024年第1期24卷 975-983页

作者： Syu, Jia-Hao Lin, Jerry Chun-Wei Biernacki, Piotr Ziebinski, Adam National Taiwan University Department of Computer Science and Information Engineering Taipei10617 Taiwan Silesian University of Technology Faculty of Automatic Control Electronics and Computer Science Department of Distributed Systems and Informatics Devices Gliwice44-100 Poland Silesian University of Technology Department of Distributed Systems and Informatic Devices Gliwice44-100 Poland

Sensors are the foundation to facilitate smart cities, smart grids, and smart transportation, and distance sensors are especially important for sensing the environment and gathering information. Researchers have developed various methods to calibrate distance sensors;however, they mainly focus on using external information for calibration, which is inefficient and computationally expensive. To alleviate the above problems, we propose machine learning (ML)-based calibration approaches for distance sensors. We develop two calibration mechanisms, where the naive mechanism directly predicts the true distance, while the residual mechanism predicts the residual for calibration. The conventional ML regression algorithms are then adopted to predict the true distance or calibration residual. Experimental results show that the designed residual calibration mechanism outperforms the naive mechanism by reducing the root mean square error (RMSE) by 92.2% and mean absolute error (MAE) by 93.3% in the fitting task. In prediction, the proposed 2-D-residual mechanism achieves an improvement of 78.6% and 65.7% compared to the state-of-the-art method. The experimental results show the effectiveness and generalization of the proposed ML-based calibration method, especially for the 2-D mechanism, which achieves the highest robustness in all fitting and prediction tasks. © 2001-2012 IEEE.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Real-Time Notifier for Blockchain Events 22

Real-Time Notifier for Blockchain Events

引用

22nd RoEduNet Conference: Networking in Education and Research, RoEduNet 2023

作者： Tanase-Paris, Cristian Carabas, Costin Tapus, Nicolae University Politehnica of Bucharest Faculty of Automatic Control and Computers Computer Science and Engineering Department Bucharest Romania

ISBN: (纸本)9798350393651

Blockchain technology gained much traction in the last few years. These decentralized databases offer security, immutability, and scalability across various applications. Decentralized applications generate vast amounts of data, known as events, that are recorded on the blockchain and are public to anyone. Some people may see opportunities for financial gains in these events and would like to know when they occur. This paper proposes a solution to process and deliver those events as real-time alerts to the users. It uses existing technologies such as message queues, multithreading, and asynchronous processing and integrates them into a scalable architecture. The results we achieved in this paper show that for an evenly distributed network traffic, which does not entirely consists of transaction bursts, the proposed solution offers reliability, efficiency, and a suitable delivery time to those wishing to integrate it into their projects. With time, this solution, or improved architectures, may form the basis of the following big-data architectures for processing blockchain events. © 2023 IEEE.

关键词： Blockchain

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：