检索结果-内蒙古大学图书馆

Moment Centralization-Based Gradient Descent Optimizers for Convolutional Neural Networks 1

International Conference on computer vision and Machine Intelligence, CVMI 2022

作者： Sadu, Sumanth Dubey, Shiv Ram Sreeja, S.R. Computer Vision Group Department of Computer Science and Engineering Indian Institute of Information Technology Andhra Pradesh Sri City India Computer Vision and Biometrics Laboratory Indian Institute of Information Technology Uttar Pradesh Allahabad India Department of Computer Science and Engineering Indian Institute of Information Technology Andhra Pradesh Sri City India

ISBN: (数字)9789811978678

ISBN: (纸本)9789811978661

Convolutional neural networks (CNNs) have shown very appealing performance for many computer vision applications. The training of CNNs is generally performed using stochastic gradient descent (SGD)-based optimization techniques. The adaptive momentum-based SGD optimizers are the recent trends. However, the existing optimizers are not able to maintain a zero mean in the first order moment and struggle with optimization. In this paper, we propose a moment centralization-based SGD optimizer for CNNs. Specifically, we impose the zero mean constraints on the first order moment explicitly. The proposed moment centralization is generic in nature and can be integrated with any of the existing adaptive momentum-based optimizers. The proposed idea is tested with three state-of-the-art optimization techniques, including Adam, Radam, and Adabelief on benchmark CIFAR10, CIFAR100, and TinyImageNet datasets for image classification. The performance of the existing optimizers is generally improved when integrated with the proposed moment centralization. Further, the results of the proposed moment centralization are also better than the existing gradient centralization. The analytical analysis using the toy example shows that the proposed method leads to a shorter and smoother optimization trajectory. The source code is made publicly available at https://***/sumanthsadhu/MC-optimizer. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

Artificial visual perception neural system using a solution-processable MoS_(2)-based in-memory light sensor

引用

Light(science & Applications) 2023年第8期12卷 1583-1594页

作者： Dayanand Kumar Lana Joharji Hanrui Li Ayman Rezk Ammar Nayfeh Nazek El-Atab Smart Advanced Memory Devices and Applications(SAMA)LaboratoryElectrical and Computer Engineering ProgramComputer Electrical Mathematical Science and Engineering DivisionKing Abdullah University of Science and Technology(KAUST)Thuwal 23955Kingdom of Saudi Arabia Department of Electrical Engineering and Computer Science Khalifa UniversityAbu Dhabi 127788United Arab Emirates

Optoelectronic devices are advantageous in in-memory light sensing for visual information processing,recognition,and storage in an energy-efficient ***,in-memory light sensors have been proposed to improve the energy,area,and time efficiencies of neuromorphic computing *** study is primarily focused on the development of a single sensing-storage-processing node based on a two-terminal solution-processable MoS2 metal-oxide-semiconductor(MOS)charge-trapping memory structure—the basic structure for charge-coupled devices(CCD)—and showing its suitability for in-memory light sensing and artificial visual *** memory window of the device increased from 2.8 V to more than 6V when the device was irradiated with optical lights of different wavelengths during the program ***,the charge retention capability of the device at a high temperature(100 ℃)was enhanced from 36 to 64%when exposed to a light wavelength of 400 *** larger shift in the threshold voltage with an increasing operating voltage confirmed that more charges were trapped at the Al_(2)O_(3)/MoS_(2) interface and in the MoS_(2) layer.A small convolutional neural network was proposed to measure the optical sensing and electrical programming abilities of the *** array simulation received optical images transmitted using a blue light wavelength and performed inference computation to process and recognize the images with 91%*** study is a significant step toward the development of optoelectronic MOS memory devices for neuromorphic visual perception,adaptive parallel processing networks for in-memory light sensing,and smart CCD cameras with artificial visual perception capabilities.

关键词： visual neural process

来源：评论

学校读者我要写书评

暂无评论

Reward Maximization for Disaster Zone Monitoring with Heterogeneous UAVs

引用

IEEE/ACM Transactions on Networking 2024年第1期32卷 890-903页

作者： Xu, Wenzheng Wang, Chengxi Xie, Hongbin Liang, Weifa Dai, Haipeng Xu, Zichuan Wang, Ziming Guo, Bing Das, Sajal K. Sichuan University College of Computer Science Chengdu610065 China City University of Hong Kong Department of Computer Science Hong Kong Hong Kong Nanjing University Department of Computer Science and Technology Nanjing210023 China Dalian University of Technology School of Software Dalian116024 China Sichuan University Key Laboratory of Birth Defects and Related Maternal and Child Diseases West China Second Hospital College of Computer Science Chengdu610066 China Missouri University of Science and Technology Department of Computer Science RollaMO65409 United States

In this paper, we study the deployment of K heterogeneous UAVs to monitor Points of Interest (PoIs) in a disaster zone, where a PoI may represent a school building or an office building, in which people are trapped. A UAV can take images/videos of PoIs and send its collected information back to a nearby rescue station for decision-making. Unlike most existing studies that focused on only homogeneous UAVs, we here study the scheduling of K heterogeneous UAVs, where different UAVs have different energy capacities and functionalities that lead to different monitoring qualities (monitoring rewards) of each PoI. For example, one type of UAVs can take only visual images while the other type of UAVs can take both visual and thermal infrared images. In this paper, we investigate a problem of scheduling K heterogeneous UAVs to monitor PoIs so that the sum of monitoring rewards received by all UAVs is maximized, subject to energy capacity on each UAV. We propose the very first 13 -approximation algorithm for this scheduling problem. We also evaluate the performance of the proposed algorithm, using real parameters of commercial UAVs. Experimental results show that the performance of the proposed algorithm is promising, which is improved by 25%, compared with existing algorithms. © 1993-2012 IEEE.

关键词： Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Application-Oriented Cloud Workload Prediction:A Survey and New Perspectives

引用

Tsinghua science and Technology 2025年第1期30卷 34-54页

作者： Binbin Feng Zhijun Ding Key Laboratory of Embedded System and Service Computing Ministry of EducationDepartment of Computer Science and TechnologyTongji UniversityShanghai 201804China Key Laboratory of Embedded System and Service Computing Ministry of EducationTongji UniversityDepartment of Computer Science and TechnologyTongji UniversityShanghai 201804China Shanghai Artificial Intelligence Laboratory Shanghai 200232China

Workload prediction is critical in enabling proactive resource management of cloud *** workload prediction is valuable for cloud users and providers as it can effectively guide many practices,such as performance assurance,cost reduction,and energy consumption ***,cloud workload prediction is highly challenging due to the complexity and dynamics of workloads,and various solutions have been proposed to enhance the prediction *** paper aims to provide an in-depth understanding and categorization of existing solutions through extensive literature *** existing surveys,for the first time,we comprehensively sort out and analyze the development landscape of workload prediction from a new perspective,i.e.,application-oriented rather than prediction methodologies per ***,we first introduce the basic features of workload prediction,and then analyze and categorize existing efforts based on two significant characteristics of cloud applications:variability and ***,we also investigate how workload prediction is applied to resource ***,open research opportunities in workload prediction are highlighted to foster further advancements.

关键词： cloud computing workload prediction resource management artificial intelligence for IT operations(AIOps)

来源：评论

学校读者我要写书评

暂无评论

FCA-based θ-iceberg core decomposition in graphs

引用

Journal of Ambient Intelligence and Humanized Computing 2024年第2期15卷 1423-1428页

作者： Hao, Fei Xinchang, Khamphaphone Park, Doo-Soon Key Laboratory of Modern Teaching Technology Ministry of Education Xi’an China School of Computer Science Shaanxi Normal University Xi’an710119 China Department of Computer Science and Engineering Soonchunhyang University Asan31538 Korea Republic of Department of Computer Software Engineering Soonchunhyang University Asan31538 Korea Republic of

Complex networking analysis is a powerful technique for understanding both complex networks and big graphs in ubiquitous computing. Particularly, there are several novel metrics, such as k-clique and k-core are proposed in order to study the relative importance of nodes in complex networks. Among of those metrics, k-core analysis is an effective approach for simplifying graphical structure. However, the relation between k and the scale of networks is not explored in most existing literature. Toward this end, this paper formulate a new research problem, θ-Iceberg Core decomposition in graphs, which is able to incorporate a parameter θ (0 © Springer-Verlag GmbH Germany, part of Springer Nature 2017.

关键词： Formal concept analysis

来源：评论

学校读者我要写书评

暂无评论

SEDARU-net: a squeeze-excitation dilated based residual U-Net with attention mechanism for automatic melanoma lesion segmentation

引用

Multimedia Tools and Applications 2024年 1-28页

作者： Lafraxo, Samira El Ansari, Mohamed Koutti, Lahcen Kerkaou, Zakaria Souaidi, Meryem LabSIV Department of Computer Science Faculty of Sciences Ibnou Zohr University Agadir80000 Morocco Department of Computer Sciences Faculty of Sciences Informatics and Applications Laboratory Moulay Ismail University Meknès Morocco

One of the most dangerous types of skin cancer, malignant melanoma, must be detected early on in order to receive successful therapy. If melanoma is not diagnosed in a timely manner, it might possibly result in death. According to clinical research, because of their varied hue, texture, and imperceptible borders, these early melanoma indications are very challenging for dermatologists to recognize. Therefore, it’s crucial to suggest an automated method that can accurately identify and distinguish between benign and malignant melanoma. Numerous automated methods have been developed by scientists to segment abnormalities from dermoscopic images. On the other hand, conventional models could find it difficult to reliably capture the multi-scale properties, which could result in inconsistent segmentation performance for a variety of object shapes and sizes. Furthermore, models with complicated forms and bounds, such as U-Net and DeepLabV3+, have difficulty properly segmenting tiny, thin, or complex lesions. Thus, we introduce A Squeeze-Excitation Dilated Residual U-Net with Attention Mechanism (SEDARU-Net) in this paper, a novel and automated semantic segmentation network for efficient skin lesion segmentation. To keep spatial information across layers and capture both local and global context, the model is built on U-net combined with dilated convolution. To solve optimization problems, residual blocks are used instead of the basic U-net units. This enhances feature learning, encourages better feature reuse, and allows for the creation of deeper and more robust networks. In order to encourage feature recalibration, global context awareness, and spatial adaptation, each residual block is supplemented with squeeze and excitation units. In addition, The attention gate is also included in the skip connection part of the network to enhance the beneficial channel dimension characteristics and suppress the unreliable background features. According to the results of the experim

关键词： Dermatology

来源：评论

学校读者我要写书评

暂无评论

Term-Modal Deontic Logics as Fragments of Alethic First-Order Modal Logics

Journal of Applied Logics

引用

Journal of Applied Logics 2025年第2期12卷 159-188页

作者： Frijters, Stef Van den Eede, Robbe Center for Logic and Philosophy of Science KU Leuven Belgium Department of Computer Science KU Leuven Artificial Intel ligence Laboratory Vrije Universiteit Brussel Belgium

In contrast to what has become known as standard deontic logic, Anderson and Kanger independently proposed alethic modal logics with a deontic constant, in which an obligation operator can be defined. It can be proven that several deontic logics are fragments of these alethic modal logics. This has been called "one of the main mathematical results on propositional monadic deontic logic". We show that a similar translation theorem can be proven for a number of predicative deontic logics, known as term-modal deontic logics (TMDLs). In this paper we define several TMDLs and several first-order (alethic) modal logics (FOMLs) with predicative constants, as well as a translation from the TMDLs to the FOMLs. We prove that a number of the TMDLs are fragments of the FOMLs. This is interesting not only from a conceptual, but also from a technical perspective. It shows that (some) term-modal logics are actually reducible to orthodox (i.e. not term-modal) FOMLs. We also present a few negative results, showing that some TMDLs cannot straightforwardly be shown to be fragments of FOMLs. Finally, we briefly discuss the possible repercussions of both the positive and negative results for other reductions, in particular for the reduction of propositional multi-modal logics to propositional uni-modal logics. © 2025, College Publications. All rights reserved.

关键词： Formal logic

来源：评论

学校读者我要写书评

暂无评论

Federated Learning-Based Approach for Effective Educational Resource Recommendation

Federated Learning-Based Approach for Effective Educational ...

引用

2024 International Conference on Advances in Electrical and Communication Technologies, ICAECOT 2024

作者： Cheriet, Nadhir Belazoui, Abdelouahab University of Batna 2 Department of Computer Science Batna Algeria University of Batna 2 Lamie Laboratory Department of Computer Science Batna Algeria

ISBN: (纸本)9798350353754

In recent years, technology-enhanced learning has affected the whole educational process by introducing innovative learner-centric tools;among them are educational recommender systems. It helps to access educational resources on the web in a flexible, suggestion-based way. This paper focuses on improving the accuracy of recommendations by proposing a novel hybrid approach based on the aggregation of several techniques and models. The proposed approach is grounded in the federated learning model, which involves replicating servers running different recommendation techniques and setting up a coordination mechanism to achieve consensus among them. Finally, it's obvious that the proposed approach holds promise for overcoming the main recommendation hurdles, such as cold starts and data sparsity, and consequently improving its accuracy. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

SDSL: Safe Drive Safe Live Application for IoV 4

SDSL: Safe Drive Safe Live Application for IoV

引用

4th International Conference on Embedded and Distributed Systems, EDiS 2024

作者： Djaidi, Malika Zerrouki, Hayet Bentitraoui, Mohamed Moncef University of Sciences and Technology Department of Computer Science Algiers Algeria University of Sciences and Technology Department of Computer Science Riima Laboratory Algiers Algeria

ISBN: (纸本)9798331519872

The Internet of Vehicles can include systems that monitor the health of drivers in real-time, detect signs of fatigue or impairment, and provide alerts to ensure safe driving conditions. With the increase in road accidents, attention is given to healthcare management in the IoV. In addition, the health of individuals, particularly drivers, can be the cause of a road accident. This is why we are interested in the detection of anomalies, in particular cardiovascular problems. In this paper, we propose a new system called Safe Drive Safe Live (SDSL) that is able to detect healthcare anomalies in any vehicle and manage them rapidly. The purpose of SDSL is to save drivers' lives and avoid accidents. The efficiency of the proposed system is evaluated through various simulations using a real map. The simulation results show that the proposed scheme is effective, useful, and feasible. © 2024 IEEE.

关键词： Highway accidents

来源：评论

学校读者我要写书评

暂无评论

A Method based on Multi-agent Systems and Passive Replication Technique for Predicting Failures in Cloud Computing

引用

Recent Advances in computer science and Communications 2023年第1期16卷 18-32页

作者： Hioual, Ouided Hioual, Ouassila Hemam, Sofiane Mounine Maifi, Lyes Mathematics and Computer Science Department Science and Technology Abbes Laghrour University Khenchela Algeria Mathematics and Computer Science Department Science and Technology Abbes Laghrour University Lire laboratory of Constantine 2 Khenchela Algeria Algiers Algeria

Background: Cloud computing refers to the computing capacities of remote computers, where the user has considerable computing power without having to own power units. The probability of failures, which can occur during execution, increases in the number of nodes. Since failures cannot be completely avoided, one solution is to use failure tolerance mechanisms. Predicting failures has become a major task for engineers and software developers, as failure increases resource usage costs. Objectives and Methods: This paper presents a hybrid method of predicting failures in a cloudcomputing environment based on the passive replication technique and multi-agent systems. It detects failures, improves the average response time and minimizes lost time. This method makes it possible to efficiently and transparently guarantee the continuity of cloud computing services in the presence of failures. Results: The results show that the proposed method performs well in the presence of failures, improves the response time and minimizes the additional costs caused by the failures. Conclusion: This paper proposes a hybrid method of predicting failures in cloud-computing based on the passive replication technique and multi-agent systems to detect failures and minimize lost time. The replication technique works by duplicating some system components, which are deployed simultaneously across different resources. This technique aims to make the system robust, increase availability and guarantee the execution of jobs. In addition, it is suitable for long-running tasks. © 2023 Bentham science Publishers.

关键词： Multi agent systems

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：