检索结果-内蒙古大学图书馆

Carrying Out CNN Channel Pruning in a White Box

IEEE Transactions on Neural Networks and Learning Systems 2023年第10期34卷 7946-7955页

作者： Zhang, Yuxin Lin, Mingbao Lin, Chia-Wen Chen, Jie Wu, Yongjian Tian, Yonghong Ji, Rongrong Xiamen University Media Analytics and Computing Laboratory Department of Artificial Intelligence School of Informatics Xiamen361005 China Youtu Laboratory Tencent Shanghai200233 China National Tsing Hua University Department of Electrical Engineering The Institute of Communications Engineering Hsinchu30013 Taiwan Peng Cheng Laboratory Shenzhen518000 China Peking University Department of Computer Science and Technology Beijing100871 China School of Informatics and the Institute of Artificial Intelligence Xiamen University Media Analytics and Computing Laboratory Department of Artificial Intelligence Xiamen361005 China

Channel pruning has been long studied to compress convolutional neural networks (CNNs), which significantly reduces the overall computation. Prior works implement channel pruning in an unexplainable manner, which tends to reduce the final classification errors while failing to consider the internal influence of each channel. In this article, we conduct channel pruning in a white box. Through deep visualization of feature maps activated by different channels, we observe that different channels have a varying contribution to different categories in image classification. Inspired by this, we choose to preserve channels contributing to most categories. Specifically, to model the contribution of each channel to differentiating categories, we develop a class-wise mask for each channel, implemented in a dynamic training manner with respect to the input image's category. On the basis of the learned class-wise mask, we perform a global voting mechanism to remove channels with less category discrimination. Lastly, a fine-tuning process is conducted to recover the performance of the pruned model. To our best knowledge, it is the first time that CNN interpretability theory is considered to guide channel pruning. Extensive experiments on representative image classification tasks demonstrate the superiority of our White-Box over many state-of-the-arts (SOTAs). For instance, on CIFAR-10, it reduces 65.23% floating point operations per seconds (FLOPs) with even 0.62% accuracy improvement for ResNet-110. On ILSVRC-2012, White-Box achieves a 45.6% FLOP reduction with only a small loss of 0.83% in the top-1 accuracy for ResNet-50. Code is available at https://***/zyxxμWhite-Box. © 2012 IEEE.

关键词： Visualization

来源：评论

学校读者我要写书评

暂无评论

Preliminary Study on the Reproducibility of Fix Templates in Static Analysis Tool 35

Preliminary Study on the Reproducibility of Fix Templates in...

引用

35th IEEE International Conference on Software engineering Education and Training, CSEE and T 2023

作者： Kim, Sohyun Kim, Youngkyoung Lee, Eunseok Sungkyunkwan University Department of Software Suwon Korea Republic of Sungkyunkwan University Department of Electrical and Computer Engineering Suwon Korea Republic of Sungkyunkwan University College of Computing and Informatics Suwon Korea Republic of

ISBN: (纸本)9798350322026

Automated Program Repair (APR) automatically generates patches for identified defects. As a result, APR can encourage novice students to learn coding by providing appropriate patches. Since students have their coding conventions, we should be able to deal with many types of defects. Existing studies have used predefined RuleId-driven templates to automatically fix defects in Static Analysis Tools(SATs). However, the community periodically adds, deletes, or changes SAT's RuleIds. This is difficult for existing RuleId-based templates to reflect those changes immediately. Existing studies only cover about 10 RuleIds, making it difficult to address all defects faced by all students. Therefore, it is necessary to establish appropriate criteria for classifying templates. The SAT has a predefined format for how Error Messages are written, and since Error Messages contain fixing actions, defects with similar Error Messages tend to have similar fixing actions. These characteristics of the Error Message are suitable for reproducible template classification criteria. Our preliminary study demonstrated that by classifying patterns based on Error Messages, we could effectively address various defects, including those in different programming languages, using a single template. This means that if a newly added RuleId corresponds to an Error Message format already in the predefined Error Message-based template, it can be modified without additional effort. We plan to construct reproducible templates for each Error Message and provide ongoing patching of defects to students. © 2023 IEEE.

关键词： Defects

来源：评论

学校读者我要写书评

暂无评论

Diffusion Model for Camouflaged Object Detection 26

Diffusion Model for Camouflaged Object Detection

引用

26th European Conference on Artificial Intelligence, ECAI 2023

作者： Chen, Zhennan Gao, Rongrong Xiang, Tian-Zhu Lin, Fan School of Informatics Xiamen University Xiamen China Department of Computer Science and Engineering Hkust Hong Kong G42 Abu Dhabi United Arab Emirates

ISBN: (纸本)9781643684369

Camouflaged object detection is a challenging task that aims to identify objects that are highly similar to their background. Due to the powerful noise-to-image denoising capability of denoising diffusion models, in this paper, we propose a diffusion-based framework for camouflaged object detection, termed diffCOD, a new framework that considers the camouflaged object segmentation task as a denoising diffusion process from noisy masks to object masks. Specifically, the object mask diffuses from the ground-truth masks to a random distribution, and the designed model learns to reverse this noising process. To strengthen the denoising learning, the input image prior is encoded and integrated into the denoising diffusion model to guide the diffusion process. Furthermore, we design an injection attention module (IAM) to interact conditional semantic features extracted from the image with the diffusion noise embedding via the cross-attention mechanism to enhance denoising learning. Extensive experiments on four widely used COD benchmark datasets demonstrate that the proposed method achieves favorable performance compared to the existing 11 state-of-the-art methods, especially in the detailed texture segmentation of camouflaged objects. Our code will be made publicly available at: https://***/ZNan-Chen/diffCOD. © 2023 The Authors.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Password Manager with Multi-Factor Authentication 1

Password Manager with Multi-Factor Authentication

引用

1st IEEE International Conference on Networking and Communications, ICNWC 2023

作者： Dhanalakshmi, R. Vijayaraghavan, N. Narasimhan, S. Basha, Saleem Vellore Institute of Technology School of Computer Science and Engineering Chennai India KCG College of Technology Department of Mathematics Chennai India Mazoon University College Computing and Informatics Department muscat Oman

ISBN: (纸本)9798350336009

Data breach is a serious issue as it leaks the personal information of more than billions of users and their privacy is compromised. More than 77% of organizations do not have a Cyber Security Incident Response plan. So, it is necessary to be informed of network security and ways to store and generate secured passwords. Having different and random passwords for ones digital accounts can exponentially increase the security of user's data. The goal of this project is to build a password manager which can securely store and encrypt passwords and other data. The multifactor authentication will provide increased security to validate the user into password management system. Multifactor authentication system includes physical security key and graphical password authentication. © 2023 IEEE.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Bibliometric and Content Analysis of Large Language Models Research in Software engineering: The Potential and Limitation in Software engineering

引用

International Journal of Advanced computer Science and Applications 2025年第4期16卷 344-356页

作者： Damayanti, Annisa Dwi Gani, Hamdan Zhipeng, Feng Gani, Helmy Zuhriyah, Sitti Nurani Djabir, St. Nurhayati Wardani, Nur Ilmiyanti Department of Environmental Engineering Faculty of Engineering Hasanuddin University Makassar Indonesia Department of Machinery Automation System ATI Makassar Polytechnic Makassar Indonesia School of Culture Creativity and Media Hangzhou Normal University Zhejiang Hangzhou China Department of Industrial Hygiene Faculty of Public Health Occupational Health and Safety Makassar College of Health Sciences Indonesia Department of Computer System Universitas Handayani Makassar Makassar Indonesia Department of Information Systems and Technology Institut Teknologi dan Bisnis Nobel Indonesia Jl. Sultan Alauddin No.212 Makassar90221 Indonesia Department of Informatics Engineering Universitas Handayani Makassar Makassar Indonesia

Large Language Models (LLM) is a type of artificial neural network that excels at language-related tasks. The advantages and disadvantages of using LLM in software engineering are still being debated, but it is a tool that can be utilized in software engineering. This study aimed to analyze LLM studies in software engineering using bibliometric and content analysis. The study data were retrieved from Web of Science and Scopus. The data were analyzed using two popular bibliometric approaches: bibliometric and content analysis. VOS Viewer and Bibliometrix software were used to conduct the bibliometric analysis. The bibliometric analysis was performed using science mapping and performance analysis approaches. Various bibliometric data, including the most frequently referenced publications, journals, and nations, were evaluated and presented. Then, the synthetic knowledge method was utilized for content analysis. This study examined 235 papers, with 836 authors contributing. The publications were published in 123 different journals. The average number of citations per publication is 1.44. Most publications were published in Proceedings International Conference on Software engineering and ACM International Conference Proceeding Series, with China and the United States emerging as the leading countries. It was discovered that international collaboration on the issue was inadequate. The most often used keywords in the publications were "software design," "code (symbols)," and "code generation." Following the content analysis, three themes emerged: 1) Integration of LLM into software engineering education, 2) application of LLM in software engineering, and 3) potential and limitation of LLM in software engineering. The results of this study are expected to provide researchers and academics with insights into the current state of LLM in software engineering research, allowing them to develop future conclusions. © (2025), (Science and Information Organization). All Rights Res

关键词： Software design

来源：评论

学校读者我要写书评

暂无评论

A Review of the Method of Auxiliary Sources (MAS) in Applied Electromagnetics 25

A Review of the Method of Auxiliary Sources (MAS) in Applied...

引用

25th International Conference on Electromagnetics in Advanced Applications, ICEAA 2024

作者： Anastassiu, Hristos T. Papakanellos, Panagiotis J. Tsitsas, Nikolaos L. Department of Computer Informatics and Telecommunications Engineering International Hellenic University End of Magnisias Str. SerresGR-62124 Greece Department of Aeronautical Sciences Hellenic Air Force Academy Dekeleia Air Force Base Dekeleia AthensGR-13671 Greece School of Informatics Aristotle University of Thessaloniki ThessalonikiGR-54124 Greece

ISBN: (纸本)9798350360974

The Method of Auxiliary Sources (MAS) is a broadly used numerical technique, purposely designed for the solution of elliptic boundary value problems in Physics and engineering, including Electromagnetics (EM). Superficially it resembles surface integral equation formulations, such as the Method of Moments (MoM), however it has been developed on different mathematical concepts. It was originally proposed by researchers in the former Soviet Union (currently Republic of Georgia) and was further refined at the National Technical University of Athens (NTUA), Greece, through an efficient Greek-Georgian collaboration in the 1990's. The fundamental concept of auxiliary sources, i.e. simulation of E M propagation behaviour in terms of fictitious, elementary fields has prompted several, unrelated scholars to develop practically the same, or very similar techniques, known in the literature by various names, e.g. "The Multifilament Current Model Method" or "The Discrete Singularity Method". Furthermore, MAS is closely related to the "Generalized Multipole Technique" (GMT) or "Multiple Multipole Technique" (MMT). The key mathematical and physical basis, common to all aforementioned methods is the following: the solution to the E M radiation or scattering problem is constructed in terms of discrete fictitious currents (the " auxiliary sources" - ASs), located away from physical boundaries, as opposed to continuous equivalent currents placed on the interfaces, utilized in MoM, for example. © 2024 IEEE.

关键词： Integral equations

来源：评论

学校读者我要写书评

暂无评论

Secure mmWave-NOMA Multi-BS Vehicular Communications Using Cooperative Jamming

Secure mmWave-NOMA Multi-BS Vehicular Communications Using C...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Yan, Yiting Ju, Ying Hu, Jiawei Liu, Lei Pei, Qingqi Yau, Kok-Lim Alvin Wu, Celimuge Zhang, Ning Xidian University School of Telecommunications Engineering Xi'an710071 China Universiti Tunku Abdul Rahman School of Internet Engineering and Computer Science Selangor 43200 Malaysia The University of Electro-Communications Graduate School of Informatics and Engineering Tokyo182-8585 Japan University of Windsor Department of Electrical and Computer Engineering WindsorN9B-3P4 Canada

ISBN: (纸本)9798350351255

The fronthaul network architecture is the key to dealing with the massive traffic effectively and providing high-quality service, and the multiple base stations (BSs) deployed by it face the gigantic data transmission, which has given the demand for high-capacity communication and information security in the vehicular network. In this paper, we combine the millimeter wave (mmWave) communication and non-orthogonal multiple access (NOMA) technologies to escalate the communication capacity of multiple vehicle users (VUs), and propose a blockage-based cooperative jamming strategy to solve potential security risks in the vehicular network. In particular, with the help of jam-mers selected by this strategy, transmission security is enhanced simultaneously without escalating the instability of connections caused by the time-varying nature of vehicular networks under the NOMA transmission mechanism when the base station (BS) does not fully understand the channel state information (CSI) of VUs. Then we comprehensively analyze the specific distribution of roadways and the distance distribution of VUs under the NOMA strategy, and derive the performance metrics of the network based on the stochastic geometry method. Numerical results show that the proposed cooperative jamming scheme can effectively improve the secrecy performance of the vehicular network. © 2024 IEEE.

关键词： Jamming

来源：评论

学校读者我要写书评

暂无评论

Blockchain-based Security for Cloud Data Storage 9

Blockchain-based Security for Cloud Data Storage

引用

9th IEEE International Conference on Smart Instrumentation, Measurement and Applications, ICSIMA 2023

作者： Bin Idrus, Mohammad Ariff Zakwan Rahman, Farah Diyana Abdul Khalifa, Othman Omran Yusoff, Nelidya Md Islamic University Malaysia Department of Electrical and Computer Engineering International Kuala Lumpur Malaysia Universiti Teknologi Malaysia Razak Faculty of Technology and Informatics Kuala Lumpur Malaysia

ISBN: (纸本)9798350343380

Data Cloud Storage is a service that allows data to save data on offsite storage system and managed by third-party and is made accessible by a web services user. It has become one of the most convenient and efficient methods to store data online. However, like other technologies and services, cloud storage has its advantages and disadvantages, but the challenge of security is the big issue. In this project, an investigation of the use of blockchain technology to improve the security of cloud data storage. The focus to address in this project is the vulnerability of cloud data storage systems to cyber-attacks and data breaches. The design and implementation of a proof-of-concept blockchain-based data storage system, and an evaluation of the system's security and performance will be presented. A proposal system to secure cloud storage could be an alternative for traditional methods. It is proposed to use blockchain as a data protection system. © 2023 IEEE.

关键词： Cloud storage

来源：评论

学校读者我要写书评

暂无评论

An Efficient and Lightweight Commitment Scheme for IoT Data Streams 19

An Efficient and Lightweight Commitment Scheme for IoT Data ...

引用

19th Annual International Conference on Distributed Computing in Smart Systems and the Internet of Things, DCOSS-IoT 2023

作者： Katsika, Angeliki Papageorgiou, Konstantinos Fakis, Alexandros Plagianakos, Vassilis Spathoulas, Georgios University of Thessaly Department of Computer Science and Biomedical Informatics Lamia Greece University of the Aegean Department of Information and Communication Systems Engineering Samos Greece Department of Information Security and Communication Technology Gjøvik Norway

ISBN: (纸本)9798350346497

Data integrity is a crucial property for offering Internet of Things (IoT) services. Usually, this is achieved by assuming a high level of trust in the service provider, a practice that induces negative implications. In the present paper, an efficient and lightweight commitment scheme for IoT data streams is presented. The proposed scheme combines KZG polynomial commitments and a blockchain smart contract to enable full data integrity while requiring trust only at the IoT device level. Given that the IoT device operates properly, no tampering can happen at any other point of the IoT workflow, without being noticed, while the protocol offers the option for automated settlement of cases where the operator tampers with users' data. A number of experiments have been conducted to assess the computation/gas fees requirements of the scheme and validate its real-world applicability. © 2023 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

A Logical Investigation of Stock Market Prediction and Analysis using Supervised Machine Learning Algorithm 1

A Logical Investigation of Stock Market Prediction and Analy...

引用

1st IEEE International Conference on Networking and Communications, ICNWC 2023

作者： Dhanalakshmi, R. Kumar, V. Vasantha Basha, Saleem Vijayaraghavan, N. Vellore Institute of Technology School of Computer Science and Engineering Chennai India Mazoon University College Computing and Informatics Department Muscat Oman KCG College of Technology Department of CSE Chennai India

ISBN: (纸本)9798350336009

In the Field of computer science, artificial intelligence (AI) is a broad field, which is concerned with structuring smart products and machines able to perform tasks which require the intellectual capability of humans. Artificial Intelligence is in need for Stock trading's future. Robotic advisers examine millions of pieces of data using artificial intelligence and carry out the analysis at the best possible price. Market forecasts are more precise and effective thanks to the analysts. For providing higher returns, the trading firms efficiently mitigate the risk using Artificial Intelligence Techniques. Our proposed Stock Market system is a Web Application which uses the Random Forest Algorithm. The Random Forest Algorithm comes under Supervised Learning. With the system, one can forecast the stock prices for the subsequent days. © 2023 IEEE.

关键词： Forecasting

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：