检索结果-内蒙古大学图书馆

37th Conference on Neural Information Processing Systems, NeurIPS 2023

作者： Batatia, Ilyes Geiger, Mario Munoz, Jose Smidt, Tess Silberman, Lior Ortner, Christoph Engineering Laboratory University of Cambridge CambridgeCB2 1PZ United Kingdom Department of Chemistry ENS Paris-Saclay Université Paris-Saclay Gif-sur-Yvette91190 France Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA United States EIA University FTA Group Antioquia Colombia Department of Mathematics University of British Columbia VancouverBCV6T 1Z2 Canada

ISBN: (纸本)9781713899921

Reductive Lie Groups, such as the orthogonal groups, the Lorentz group, or the unitary groups, play essential roles across scientific fields as diverse as high energy physics, quantum mechanics, quantum chromodynamics, molecular dynamics, computer vision, and imaging. In this paper, we present a general Equivariant Neural Network architecture capable of respecting the symmetries of the finite-dimensional representations of any reductive Lie Group G. Our approach generalizes the successful ACE and MACE architectures for atomistic point clouds to any data equivariant to a reductive Lie group action. We also introduce the lie-nn software library, which provides all the necessary tools to develop and implement such general G-equivariant neural networks. It implements routines for the reduction of generic tensor products of representations into irreducible representations, making it easy to apply our architecture to a wide range of problems and groups. The generality and performance of our approach are demonstrated by applying it to the tasks of top quark decay tagging (Lorentz group) and shape recognition (orthogonal group). © 2023 Neural information processing systems foundation. All rights reserved.

关键词： Lie groups

来源：评论

学校读者我要写书评

暂无评论

Hybrid Search for Efficient Planning with Completeness Guarantees

arXiv

引用

arXiv 2023年

作者： Kujanpää, Kalle Pajarinen, Joni Ilin, Alexander Department of Computer Science Aalto University Finland Department of Electrical Engineering and Automation Aalto University Finland Finnish Center for Artificial Intelligence FCAI Finland System 2 AI

Solving complex planning problems has been a long-standing challenge in computer science. Learning-based subgoal search methods have shown promise in tackling these problems, but they often suffer from a lack of completeness guarantees, meaning that they may fail to find a solution even if one exists. In this paper, we propose an efficient approach to augment a subgoal search method to achieve completeness in discrete action spaces. Specifically, we augment the high-level search with low-level actions to execute a multi-level (hybrid) search, which we call complete subgoal search. This solution achieves the best of both worlds: the practical efficiency of high-level search and the completeness of low-level search. We apply the proposed search method to a recently proposed subgoal search algorithm and evaluate the algorithm trained on offline data on complex planning problems. We demonstrate that our complete subgoal search not only guarantees completeness but can even improve performance in terms of search expansions for instances that the high-level could solve without low-level augmentations. Our approach makes it possible to apply subgoal-level planning for systems where completeness is a critical requirement. Copyright © 2023, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Safeguarding Smart Vehicles: GNN-Powered Real- Time IDS for CAN Networks

Safeguarding Smart Vehicles: GNN-Powered Real- Time IDS for ...

引用

Machine Intelligence and Smart Innovation (ICMISI), International Conference on

作者： Sarah El-Dawy Abdallah Hussien Mostafa Ammar Ahmed El-Dawy Software Development Engineer S2S Vista Virtual Platforms Siemens EDA Cairo Egypt Electrical and Computer Engineering Department University of Waterloo Waterloo Canada Computer Network and Datacenter Department Arab Academy for Science Technology and Submarine Alexandria Egypt Electrical Power Department University of Alexandria Alexandria Egypt

ISBN: (数字)9798350365740

ISBN: (纸本)9798350365757

With the increasing integration of Controller Area Network (CAN) in modern vehicles, securing the communication network from malicious attacks is imperative. A novel Graph Neural Network (GNN)-based Intrusion Detection System (IDS) is proposed for real-time detection of several attacks, such as Denial of Service (DOS), fuzzy, and spoofing attacks. Our proposed model represents the CAN network as a graph, where nodes correspond to message arbitration ID and edges denote the communication sequence. Using Graph Attention Networks (GAT), our model captures the intricate relationships and dependencies between Electronic Circuit Units (ECUs), in real-time, enhancing its ability to discern normal behavior from malicious activity. The key to our system's success is its emphasis on achieving real-time capabilities. The GAT architecture optimizes computational efficiency for swift analysis of CAN traffic without compromising detection accuracy. Our system minimizes model complexity for enhanced deployability in resource-constrained vehicular environments. Comprehensive experiments on a diverse dataset showcase its ability to swiftly and accurately detect a range of attacks in real-time, with an overall accuracy of 99.75% for DOS attacks, 99.50% for fuzzy attacks, and 100% for spoofing attacks. Moreover, attack detection time is calculated to be 6.76 ms for DOS, 6.79 ms for fuzzy, and 6.69 ms for spoofing. These results made the model meet real-time detection constraints.

关键词： Technological innovation Accuracy Computational modeling Image edge detection Intrusion detection Real-time systems Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep multi-metric training: the need of multi-metric curve evaluation to avoid weak learning

引用

Neural Computing and Applications 2024年第30期36卷 18841-18862页

作者： Mamalakis, Michail Banerjee, Abhirup Ray, Surajit Wilkie, Craig Clayton, Richard H. Swift, Andrew J. Panoutsos, George Vorselaars, Bart Department of Psychiatry University of Cambridge Herchel Smith Building Robinson Way CambridgeCB2 0SZ United Kingdom Department of Computer Science University of Sheffield 211 Portobello SheffieldS1 4DP United Kingdom Insigneo Institute for in silico Medicine University of Sheffield The Pam Liversidge Building SheffieldS1 3JD United Kingdom School of Mathematics and Statistics University of Glasgow GlasgowG12 8QQ United Kingdom Department of Infection Immunity and Cardiovascular Disease University of Sheffield SheffieldS10 2HQ United Kingdom Department of Automatic Control and Systems Engineering University of Sheffield SheffieldS10 2TN United Kingdom School of Mathematics and Physics University of Lincoln Brayford Pool LincolnLN6 7TS United Kingdom Department of Engineering Science Institute of Biomedical Engineering University of Oxford OxfordOX3 7DQ United Kingdom Division of Cardiovascular Medicine Radcliffe Department of Medicine University of Oxford OxfordOX3 9DU United Kingdom

The development and application of artificial intelligence-based computer vision systems in medicine, environment, and industry are playing an increasingly prominent role. Hence, the need for optimal and efficient hyperparameter tuning strategies is more than crucial to deliver the highest performance of the deep learning networks in large and demanding datasets. In our study, we have developed and evaluated a new training methodology named deep multi-metric training (DMMT) for enhanced training performance. The DMMT delivers a state of robust learning for deep networks using a new important criterion of multi-metric performance evaluation. We have tested the DMMT methodology in multi-class (three, four, and ten), multi-vendors (different X-ray imaging devices), and multi-size (large, medium, and small) datasets. The validity of the DMMT methodology has been tested in three different classification problems: (i) medical disease classification, (ii) environmental classification, and (iii) ecological classification. For disease classification, we have used two large COVID-19 chest X-rays datasets, namely the BIMCV COVID-19+ and Sheffield hospital datasets. The environmental application is related to the classification of weather images in cloudy, rainy, shine or sunrise conditions. The ecological classification task involves a classification of three animal species (cat, dog, wild) and a classification of ten animals and transportation vehicles categories (CIFAR-10). We have used state-of-the-art networks of DenseNet-121, ResNet-50, VGG-16, VGG-19, and DenResCov-19 (DenRes-131) to verify that our novel methodology is applicable in a variety of different deep learning networks. To the best of our knowledge, this is the first work that proposes a training methodology to deliver robust learning, over a variety of deep learning networks and multi-field classification problems. © Crown 2024.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Fortifying Information Security: Security Implications of Microservice and Monolithic Architectures

Fortifying Information Security: Security Implications of Mi...

引用

Security of Information and Networks (SIN), International Conference on

作者： Sanghamitra Nemmini S Abhishek Anjali T Sincy Raj Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India Vrije Universiteit Brussel MACS Pleinlaan 2 Brussels Belgium

Monolithic architecture is a traditional and primary method for creating software programs. On the other hand, monolithic architecture's shortcomings were addressed by developing the unique evolutionary technique known as microservice architecture, which led to businesses' widespread acceptance of this design. However, the microservice design raises several security concerns, particularly web application security, a significant business problem. The purpose of this study is to analyse and contrast the security and safety features of monolithic and microservice architectures, emphasizing the security issues that each presents and the solutions that have been proposed to address them. Examining both architectures security risks and investigating viable mitigation strategies are covered in the paper. Microservices improve security and flexibility but also need a more complex and dynamic security strategy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

5GCIDS: An Intrusion Detection System for 5G Core with AI and Explainability Mechanisms

5GCIDS: An Intrusion Detection System for 5G Core with AI an...

引用

2023 IEEE Globecom Workshops, GC Wkshps 2023

作者： Radoglou-Grammatikis, Panagiotis Nakas, George Amponis, George Giannakidou, Sofia Lagkas, Thomas Argyriou, Vasileios Goudos, Sotirios Sarigiannidis, Panagiotis K3Y Ltd William Gladstone 31 Sofia1000 Bulgaria University of Western Macedonia Department of Electrical and Computer Engineering Campus ZEP Kozani Kozani50100 Greece International Hellenic University Department of Computer Science Kavala Campus Kavala65404 Greece Kingston University London Department of Networks and Digital Media Penrhyn Road Surrey Kingston upon ThamesKT1 2EE United Kingdom School of Physics Aristotle University of Thessaloniki Thessaloniki54124 Greece

ISBN: (纸本)9798350370218

The progression of fifth-generation (5G) networks provides multiple advantages, such as faster speed, reduced latency and increased capacity. Towards these advancements, it is clear that 5G Core (5GC) represents the heart of a 5G network, providing a variety of new network services such as Ultra-reliable low-latency communication (URLLC) and Massive machine-type communication (mMTC). However, despite the various benefits, 5GC is prone to several cyberthreats that can result in catastrophic effects. In this paper, an Intrusion Detection System (IDS) for 5GC is introduced. The proposed IDS called 5GCIDS adopts Artificial Intelligence (AI) methods in order to detect potential cyberattacks against Packet Forwarding Control Protocol (PFCP), which is utilised for the N4 interface between Session Management Function (SMF) andUser Plane Function (UPF). For the detection process, both Transmission Control Protocol (TCP)/Internet Protocol (IP) flow statistics and application-layer PFCP flow statistics are used. In the second case, we provide a bidirectional flow statistics generator called PFCPFlowMeter. Finally, the detection outcomes are explained as local and global explanations with the TreeSHAP method. The evaluation results demonstrate the efficiency of the proposed IDS. © 2023 IEEE.

关键词： 5G Artificial Intelligence Intrusion Detection PFCP

来源：评论

学校读者我要写书评

暂无评论

Decomposition gases of insulating oil in oil-impregnated paper insulation system with an oil gap - Characteristic change of decomposition gases due to partial discharge conditions

引用

IEEJ Transactions on Fundamentals and Materials 2021年第4期141卷 252-259页

作者： Makino, Yuta Kurihara, Takashi Takahashi, Toshihiro Kurimoto, Muneaki Kato, Takeyoshi Central Research Institute of Electric Power Industry 2-6-1 Nagasaka Yokosuka240-0196 Japan Department of Electrical Engineering and Computer Science Graduate School of Engineering Nagoya University Furo-cho Chikusa-ku Nagoya464-8603 Japan

Decomposition gases due to partial discharge (PD) conditions in an oil-impregnated paper insulation system with an oil gap were experimentally investigated. Sheets of oil-impregnated paper were sandwiched between parallel plane electrodes. An oil gap was made up by overlapping sheets of oil-impregnated paper with a hole. Insulating oil was taken from the oil gap after PDs were generated in the short or long duration, and the dissolved gas analysis was conducted. As a result, it was revealed that there was a difference in the ratio of each gas volume to the total volume of gases. The ratios of H2 and C2H2 were higher than those of other gases in short duration cases, while the ratios of H2, CH4 and C2H6 were higher than those of other gases in long duration cases. PDs occurred through the insulating oil in the short duration cases. On the other hand, PDs occurred in gases in the long duration cases. The result suggests that the difference of main decomposition gases in two cases is due to the difference of the place where PDs occurred. © 2021 The Institute of Electrical Engineers of Japan.

关键词： Partial discharges

来源：评论

学校读者我要写书评

暂无评论

Robots in education: Influence of Regulatory Focus Theory

Robots in education: Influence of Regulatory Focus Theory

引用

IEEE International Workshop on Robot and Human Communication (ROMAN)

作者： Xiaoxuan Hei Heng Zhang Adriana Tapus Computer Science and System Engineering Department (U2IS) Autonomous Systems and Robotics Lab ENSTA Paris Institut Polytechnique de Paris Paris France

The Covid-19 pandemic has massively developed the use of distance learning. The limits of this practice have gradually come to light, both for students and for teachers. It is now crucial to design alternative solutions to overcome the shortcomings of videoconferencing in terms of involvement, concentration, learning, and equity. Social robots are increasingly used as tutors in the educational context and help improve teaching efficiency. Many psychology-based principles have been applied in education to guide instructional strategies, motivate students, and create a positive and productive learning environment. In this work, we use Regulatory Focus Theory (RFT), which categorizes an individual’s motivation into two types: Promotion and Prevention. Promotion-focused individuals are motivated by the potential for growth and achievement, whereas prevention-focused individuals are motivated by the potential for avoiding negative outcomes. Based on RFT, we aim to explore if and how the regulatory-focused behavior of the tutor robot can affect participants’ learning outcomes. In this work, a language learning scenario was designed with two conditions: (1) a robot tutor with promotion-focused behavior, (2) a robot tutor with prevention-focused behavior. The results are encouraging and support that promotion robot tutor can increase the learning efficiency of promotion participants and prevention robot tutor will enhance the learning interest of prevention participants.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Naturalness perception of 2.5D prints: elevation and size of prints relation

Naturalness perception of 2.5D prints: elevation and size of...

引用

London Imaging Meeting, LIM 2023

作者： Kadyrova, Altynay Pedersen, Marius Westland, Stephen Weijkamp, Clemens Horiuchi, Takahiko Department of Computer Science Norwegian University of Science and Technology Gjøvik Norway School of Design University of Leeds LeedsLS2 9JT United Kingdom Canon Production Printing Netherlands B.V Venlo5914 HH Netherlands Graduate School of Engineering Chiba University Chiba Japan

ISBN: (纸本)9780892083633

Naturalness is a complex concept and a number of parameters might impact naturalness perception. In this work, we addressed how the combination of differ ent elevation levels and size of prints impacts natural ness perception of 2.5D prints. The results of a sub jective ranking experiment showed that observers per ceived 2.5D prints as more natural at higher elevation with larger size of print. Moreover, we observed that elevation seems to be a more dominant parameter than size of the print for observers when evaluating natural ness. © 2023 Society for Imaging science and Technology.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Trajectory Optimization and Beamforming Design for NOMA-enabled Multi-Objective RIS-aided IS-UAV-TNs with Deep Reinforcement Learning 23

Trajectory Optimization and Beamforming Design for NOMA-enab...

引用

23rd IEEE International Conference on Communication Technology, ICCT 2023

作者： Wu, Min Guo, Kefeng Cao, Haotong Lin, Zhi An, Kang Wang, Jiangzhou School of Space Information Space Engineering University Beijing101407 China Hong Kong Polytechnic University Department of Computing Hong Kong College of Electronic Engineering National University of Defense Technology Hefei230037 China School of Computer Science and Engineering Macau University of Science and Technology China Sixty-third Research Institute National University of Defense Technology Nanjing210007 China School of Engineering University of Kent CanterburyCT2 7NZ United Kingdom

ISBN: (纸本)9798350325959

This paper investigates the collaborative optimization performance route on the reconfigurable intelligent surface (RIS)-aided integrated satellite unmanned aerial vehicle-terrestrial network (IS-UAV-TN). It mainly has two primary challenges: long transmission paths with severe signal fading and the presence of obstacles along the wireless environment. To efficiently address above issues, it is assumed that RIS is deployed on the UAV to enable precise control over signal transmission, and non-orthogonal multiple access (NOMA) transmission strategies are employed at the transmitter. Based on the different transmission phases, a multi-vector optimization problem is proposed with the aim of maximizing the system achievable rate by optimizing the UAV path design, transmit beam and RIS passive phase shift. Especially the deep reinforcement learning (DRL)-based multi-vector deep deterministic policy gradient (MV-DDPG) algorithm, is utilized to achieve adaptive decision-making in non-deterministic wireless environments. Experimental results show that the proposed method has the ability to efficiently consider multiple optimization objectives designed compared to traditional comparison schemes. © 2023 IEEE.

关键词： Deep reinforcement learning integrated satellite-unmanned aerial vehicle-terrestrial networks (IS-UAV-TNs) multi-objective DDPG non-orthogonal multiple access re-configurable intelligent surface

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：