检索结果-内蒙古大学图书馆

13th International Physics Seminar 2024, IPS 2024

作者： Pebrianti, Dwi Kee, Wong Jack Bayuaji, Luhur Solichin, Achmad Department of Mechanical & Aerospace Engineering International Islamic University Malaysia Jl. Gombak Kuala Lumpur53100 Malaysia Faculty of Information Technology Universitas Budi Luhur Jl. Ciledug Raya Jakarta12260 Indonesia Faculty of Electrical & Electronics Engineering Universiti Malaysia Pahang Al-Sultan Abdullah Pekan26600 Malaysia Faculty of Data Science & Information Technology INTI International University Persiaran Perdana BBN Putra Nilai Nilai71800 Malaysia

Micro Aerial Vehicles (MAVs) has gained attentions since more than two decades ago starting from the applications in air combat up to civil applications such as packet deliveries, environmental monitoring, and surveillance. In an environment such as cities that grows denser, navigation and control for drones becomes challenging to ensure safe navigation around buildings and other obstacles. This study proposes an approach for obstacle avoidance for MAVs by using ultrasonic sensors. Four sensors are strategically positioned to cover the front, right, back, and left directions. Additionally, a downward-facing sensor measures the quadrotor's height above ground. Our goal is to develop autonomous MAV that can avoid obstacles, ensuring safe flight even in complex urban landscapes. The scenario implemented in the study is by introducing obstacle in any directions. When an obstacle is detected by the ultrasonic sensor, the signal will be sent to microcontroller and the attitude of the MAVs, roll or pitch will be adjusted to avoid the obstacle by moving to the counter direction of the obstacle. We conducted 20 trials of experiments by varying the gain values of Proportional Integral Derivative (PID) values, we fine-tune our obstacle avoidance algorithm. Modifications include optimizing roll and pitch adjustments, refining detection height thresholds, and implementing countermeasures after obstacle clearance. The results show that our proposed method has 10% overshoot when detecting any obstacles in different directions to avoid the obstacles. Our findings contribute to the advancement of safe and efficient urban drone operations, bridging the gap between technology and real-world challenges. © Published under licence by IOP Publishing Ltd.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

LGDB-Net: Dual-Branch Path for Building Extraction from Remote Sensing Image 30

LGDB-Net: Dual-Branch Path for Building Extraction from Remo...

引用

30th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2024

作者： Zhang, Ronghuan Zhao, Jing Li, Ming Zou, Qingzhi Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China Shandong University of Traditional Chinese Medicine School of Intelligence and Information Engineering Jinan China

ISBN: (纸本)9798331515966

Extracting buildings from remote sensing images using deep learning techniques is a widely applied and crucial task. Convolutional Neural Networks (CNNs) adopt hierarchical feature representation, showcasing powerful capabilities in extracting local information but facing challenges in capturing global features. Transformers can address this limitation, but they perform poorly in extracting local features and significantly increase memory requirements and computational complexity. To overcome these challenges, we propose a method for building extraction from remote sensing images called LGDB-Net (Local-Global Dual-Branch Network), employing a dual-branch approach. Firstly, inspired by Swin Transformer, we designed GB-Former(Global Branch-Former) as the backbone network to model global information. We use a linear multi-head self-attention mechanism to reduce time and memory complexity while maintaining a large global receptive field. Additionally, we replace the traditional multi-layer perceptron with a convolution-enhanced multi-layer perceptron to improve channel feature representation, reduce model parameters, and enhance segmentation performance. Secondly, we use multiple Depth-wise Conv3×3 + LN (Layer Normalization) + GeLU (Gaussian Error Linear Unit) modules as the auxiliary branch for local detailed feature extraction. Finally, we adopt a multi-scale feature fusion strategy to integrate feature information from both branches. We conduct a series of experiments on three datasets: WHU, Massachusetts, and Inria. The experimental results demonstrate that the proposed method not only effectively improves segmentation accuracy with lower building omission and commission rates but also significantly reduces model parameters and computational complexity. © 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

REINFORCING PRE-TRAINED MODELS USING COUNTERFACTUAL IMAGES

arXiv

引用

arXiv 2024年

作者： Li, Xiang Togo, Ren Maeda, Keisuke Ogawa, Takahiro Haseyama, Miki Graduate School of Information Science and Technology Hokkaido University Japan Faculty of Information Science and Technology Hokkaido University Japan Data-Driven Interdisciplinary Research Emergence Department Hokkaido University Japan

This paper proposes a novel framework to reinforce classification models using language-guided generated counterfactual images. Deep learning classification models are often trained using datasets that mirror real-world scenarios. In this training process, because learning is based solely on correlations with labels, there is a risk that models may learn spurious relationships, such as an overreliance on features not central to the subject, like background elements in images. However, due to the black-box nature of the decision-making process in deep learning models, identifying and addressing these vulnerabilities has been particularly challenging. We introduce a novel framework for reinforcing the classification models, which consists of a two-stage process. First, we identify model weaknesses by testing the model using the counterfactual image dataset, which is generated by perturbed image captions. Subsequently, we employ the counterfactual images as an augmented dataset to fine-tune and reinforce the classification model. Through extensive experiments on several classification models across various datasets, we revealed that fine-tuning with a small set of counterfactual images effectively strengthens the model. © 2024, CC BY.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

Hybrid Transfer Learning and Deep Learning for Cloud Intrusion Detection Using Bio-Inspired Optimization

Hybrid Transfer Learning and Deep Learning for Cloud Intrusi...

引用

Computational Intelligence Approaches and Applications (ICCIAA), International Conference on

作者： Salam Al-E'mari Yousef Sanjalawe Waleed Awwad Hamzah Alqudah Mohammed Aladaileh Information Security Department Faculty of Information Technology University of Petra (UoP) Amman Jordan Information Technology Department King Abdullah II School for Information Technology The University of Jordan (JU) Amman Jordan Department of Computer Science College of Science Northern Border University (NBU) Arar KSA Department of Data Science and Artificial Intelligence Faculty of Information Technology American University of Madaba Amman Jordan Department of Computer Science Faculty of Information Technology University of Petra (UoP) Amman Jordan

ISBN: (数字)9798331523657

ISBN: (纸本)9798331523664

This paper proposes a novel hybrid model for cloud intrusion detection, integrating transfer learning with bio-inspired optimization techniques to enhance detection accuracy and efficiency. Cloud environments present unique challenges due to their dynamic and shared-resource nature, making traditional Intrusion Detection Systems (IDS) inadequate. The proposed model leverages Convolutional Neural Networks (CNNs) for feature extraction, utilizing pre-trained models through transfer learning to reduce training time while maintaining high detection performance. In addition, bio-inspired optimization algorithms, specifically Cuckoo Search Algorithm (CSA) and Particle Swarm Optimization (PSO), are employed to fine-tune the CNN's hyperparameters, such as learning rate and batch size, ensuring optimal performance. The hybrid approach addresses the limitations of manual tuning and improves the model's ability to generalize across different cloud environments. Using the CICIDS 2017 dataset, the proposed system demonstrates superior performance in detecting both known and unknown threats, including zero-day attacks. The experimental results show significant improvements in accuracy of 99.3%, precision of 98.7%, recall of 98.5%, and an F1-score of 98.6% compared to state-of-the-art models.

关键词： Training Accuracy Biological system modeling Heuristic algorithms Transfer learning Intrusion detection Convolutional neural networks Particle swarm optimization Optimization Tuning

来源：评论

学校读者我要写书评

暂无评论

ENHANCING GENERATIVE CLASS INCREMENTAL LEARNING PERFORMANCE WITH MODEL FORGETTING APPROACH

arXiv

引用

arXiv 2024年

作者： Togo, Taro Togo, Ren Maeda, Keisuke Ogawa, Takahiro Haseyama, Miki Graduate School of Information Science and Technology Hokkaido University Japan Faculty of Information Science and Technology Hokkaido University Japan Data-Driven Interdisciplinary Research Emergence Department Hokkaido University Japan

This study presents a novel approach to Generative Class Incremental Learning (GCIL) by introducing the forgetting mechanism, aimed at dynamically managing class information for better adaptation to streaming data. GCIL is one of the hot topics in the field of computer vision, and this is considered one of the crucial tasks in society, specifically the continual learning of generative models. The ability to forget is a crucial brain function that facilitates continual learning by selectively discarding less relevant information for humans. However, in the field of machine learning models, the concept of intentionally forgetting has not been extensively investigated. In this study we aim to bridge this gap by incorporating the forgetting mechanisms into GCIL, thereby examining their impact on the models’ ability to learn in continual learning. Through our experiments, we have found that integrating the forgetting mechanisms significantly enhances the models’ performance in acquiring new knowledge, underscoring the positive role that strategic forgetting plays in the process of continual learning. Copyright © 2024, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Automating the Diagnosis of Cucumber Plant Diseases Using Machine Learning 25

Automating the Diagnosis of Cucumber Plant Diseases Using Ma...

引用

25th International Arab Conference on information technology, ACIT 2024

作者： Banimustafa, Ahmed Qattouz, Hazem Almatarneh, Sattam Mansour, Khalid Alazaidah, Raed Bulkrock, Olla Higher Colleges of Technology Computer Information Science Dept Abu Dhabi United Arab Emirates Princess Sumaya University for Technology Software Engineering Dept Amman Jordan Zarqa University Data Science & Artificial Intelligence Dept Zarqa Jordan University of Business and Technology College of Engineering Jeddah Saudi Arabia Arab-Open University Bahrain Faculty of Computer Studies Bahrain

ISBN: (纸本)9798331540012

Farmers all over the world struggle to fight plant diseases. However, the incorrect use of pesticides can waste time, effort, and money and can also be very harmful to the environment and human health. In this work, we present an empirical study that involves identifying two pathogens related to Cucumbers plant diseases: Powdery Mildew and Alternaria Cucumerina. The study aims to develop a computer-based solution based on deep learning and machine learning algorithms to detect these diseases. To achieve this, a dataset of hundreds of samples was collected from ordinary farms in Beqaa, an agricultural region 15 KM north of Amman, Jordan. Three machine learning models were created in this study using Support Vector Machines (SVM), Logistic Regression (LR), and Multi-layer perceptron algorithms. The constructed models were applied to the dataset after embedding the dataset images using the Convolutional Neural Networks (CNN) InceptionV3 pre-trained model. The model was tested and evaluated using cross-evaluation and F1-Score, CA, AUC, Precision, Recall, and Specificity, which all showed excellent results, particularly using the SVM algorithm. The constructed methods are intended to be deployed to farms in their field condition environments using a CCTV system that can capture images of plants during growth, which can help identify plant diseases at that right stage, which would save time, effort, and money and reduce the impact of wrong and overuse of pesticides on consumers and the environment. © 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

Reinforcing Pre-Trained Models Using Counterfactual Images

Reinforcing Pre-Trained Models Using Counterfactual Images

引用

IEEE International Conference on Image Processing

作者： Xiang Li Ren Togo Keisuke Maeda Takahiro Ogawa Miki Haseyama Graduate School of Information Science and Technology Hokkaido University Japan Faculty of Information Science and Technology Hokkaido University Japan Data-Driven Interdisciplinary Research Emergence Department Hokkaido University Japan

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

关键词： Deep learning Training Image recognition Decision making data augmentation Robustness data models

来源：评论

学校读者我要写书评

暂无评论

A Hybrid MTS Anomaly Detection Method Based on Reconstruction and Adaptive Spatial-Temporal Graph Network 13

A Hybrid MTS Anomaly Detection Method Based on Reconstructio...

引用

13th IEEE data Driven Control and Learning Systems Conference, DDCLS 2024

作者： Yuan, X. Ning, Ke Wei, K. He, Yan-Lin Zhu, Qun-Xiong Zhang, Ming-Qing Luo, Yi Zhang, Yang College of Information Science and Technology Beijing University of Chemical Technology Beijing100020 China Ministry of Education of China Engineering Research Center of Intelligent Pse Beijing100020 China Macao Polytechnic University Faculty of Applied Sciences 999078 China Research Institute of Mine Big Data Chinese Institute of Coal Science Beijing100013 China

ISBN: (纸本)9798350361674

Effective anomaly detection of multivariate time series data is crucial for modern industrial applications. However, existing works usually ignore spatial and temporal dependencies of series. The data with complicated spatial-temporal feature put forward higher requirements for accurate detection of anomalies. In this paper, we propose an anomaly detection method for multivariate time series based on spatial-temporal network (STN). Firstly, the time series are reconstructed by Generative Adversarial Networks (GAN). The reconstruction block is used to amplify the influence of the abnormal state and prompt the anomalies more detectable, as poorly reconstructed data of abnormal state further worsen the quality of the final predicted value. Secondly, an adaptive learning graph structure layer is proposed. The graph structure is learned by adaptive method without any prior knowledge and is used for reconstructed data prediction. Thirdly, we use STN to extract spatial and temporal dependencies of series. Especially, attention mechanism is introduced into graph convolution network to enhance the prediction performance by focusing on more important neighbor node message. Finally, our model is tested on 3 public datasets to evaluate performance. The results demonstrate that our method performs better than baselines. © 2024 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Introducing Class Replacement Technique in Class Incremental Learning in Generative Models

Introducing Class Replacement Technique in Class Incremental...

引用

IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)

作者： Taro Togo Ren Togo Keisuke Maeda Takahiro Ogawa Miki Haseyama Graduate School of Information Science and Technology Hokkaido University Japan Faculty of Information Science and Technology Hokkaido University Japan Data-Driven Interdisciplinary Research Emergence Department Hokkaido University Japan

ISBN: (数字)9798350386844

ISBN: (纸本)9798350386851

This paper presents a novel method for Class Incremental Learning (CIL) in image-generative models, using a class replacement technique. There has been little focus on unnecessary classes in CIL for image-generative models. Our approach is inspired by human forgetting and cognitive processes. By replacing unnecessary classes with noise images, we improve the performance of learning new classes in CIL. We performed fine-tuning with replacing techniques in MNIST and investigated the impact of replacing unnecessary classes. Through this study, we aim to promote the continual use of image-generative models by replacing unnecessary classes.

关键词： Continuing education Incremental learning Cognitive processes Noise Consumer electronics

来源：评论

学校读者我要写书评

暂无评论

Hybrid Agile Model, the Synergy of Waterfall Model and Scrum in a Software Project

Hybrid Agile Model, the Synergy of Waterfall Model and Scrum...

引用

information technology and Digital Applications (ICITDA), IEEE International Conference on

作者： Norzariyah Yahya Abdul Muizz Abdul Salam Siti Sarah Maidin Normi Sham Awang Abu Bakar Department of Computer Science Kulliyyah of Information and Communication Technology International Islamic University Malaysia Kuala Lumpur Malaysia Faculty of Data Science and Information Technology (FDSIT) INTI International University Malaysia

ISBN: (数字)9798331510732

ISBN: (纸本)9798331510749

Hybrid agile is one of the recent software engineering process model used by a software engineering team in a software development project. A hybrid agile model is used by software engineering team to meet the requirements of a software project. However, what does the hybrid agile model used by a software engineering team look like? A qualitative study was conducted to answer this question, and a series of interviews were chosen as a data collection instrument. The interview participants are practitioners involved in software projects using hybrid agile. The recorded interviews are transcribed and analysed using thematic analysis. The finding shows that the software engineering team uses the Waterfall model in the planning, design, deployment and maintenance phases, while scrum in the development and testing phases. The results have led to the hybrid agile model presented in this study, which leverages the strengths of each model to ensure the completion of a software project. Future iterations of this study will assess a project's software quality, comparing the proposed model with plan-driven and scrum.

关键词： Instruments Software quality Planning Maintenance Scrum (Software development) Interviews information technology Software development management Software engineering Testing

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：