检索结果-内蒙古大学图书馆

Convergence in Engineering (ICCE), ieee International conference for

作者： Sandip Paul Kumar Sankar Ray Diganta Saha Indian Statistical Institute Kolkata India Jadavpur University Kolkata India

ISBN: (数字)9781728173405

ISBN: (纸本)9781728173412

Hybrid deep neural-symbolic architecture for event-detection employs a deep neural network at the back-end to perform low-level reasoning and a symbolic logical module to perform high-level cognitive reasoning. The currently known hybrid architectures use classical Answer Set programming(ASP), which is unable to perform fuzzy reasoning with uncertainty. Moreover these systems don't extract new rules from the available data. On the other hand, there are neuro-fuzzy systems that can extract fuzzy rules from data by means of Gaussian Restricted Boltzman Machines (GRBM). Both the aspects should be merged together to achieve human-like intelligent reasoning and learning from environment. But the success of such an integration depends upon the chosen logical system, that can support fuzzy reasoning with uncertainty, as well as, can support the extracted knowledge from GRBM. This work investigates the feasibility of using interval-valued fuzzy logic programming for this purpose. This work focuses on the theoretical aspects from logic programming perspective.

关键词： Cognition Uncertainty Data mining Feature extraction Training Neural networks Computer architecture

来源：评论

学校读者我要写书评

暂无评论

Deep learning for MRI-based CT synthesis: a comparison of MRI sequences and neural network architectures

Deep learning for MRI-based CT synthesis: a comparison of MR...

引用

ieee Nuclear Science Symposium / Medical Imaging conference (NSS/MIC)

作者： Larroza, Andres Moliner, Laura Alvarez-Gomez, Juan M. Oliver, Sandra Espinos-Morato, Hector Vergara-Diaz, Marina Rodriguez-Alvarez, Maria J. Univ Politecn Valencia Inst Instrumentac Imagen Mol I3M Consejo Super Invest Cient Camino Vera S-N Valencia 46022 Spain

ISBN: (纸本)9781728141640

Synthetic computed tomography (CT) images derived from magnetic resonance images (MRI) are of interest for radiotherapy planning and positron emission tomography (PET) attenuation correction. In recent years, deep learning implementations have demonstrated improvement over atlas-based and segmentation-based methods. Nevertheless, several open questions remain to be addressed, such as which is the best of MRI sequences and neural network architectures. In this work, we compared the performance of different combinations of two common MRI sequences (T1- and T2-weighted), and three state-of-the-art neural networks designed for medical image processing (Vnet, HighRes3dNet and ScaleNet). The experiments were conducted on brain datasets from a public database. Our results suggest that T1 images perform better than T2, but the results further improve when combining both sequences. The lowest mean average error over the entire head (MAE = 101.76 +/- 10.4 HU) was achieved combining T1 and T2 scans with HighRes3dNet. All tested deep learning models achieved significantly lower MAE (p < 0.01) than a well-known atlas-based method.

关键词： Magnetic resonance

来源：评论

学校读者我要写书评

暂无评论

A Dynamic DL-Driven Architecture to Combat Sophisticated Android Malware

引用

ieee ACCESS 2020年 8卷 129600-129612页

作者： Bibi, Iram Akhunzada, Adnan Malik, Jahanzaib Iqbal, Javed Mussaddiq, Arslan Kim, Sungwon COMSATS Univ Comp Sci Dept Islamabad 46000 Pakistan Tech Univ Denmark DTU Comp DK-2800 Copenhagen Denmark NUST Natl Cyber Secur Auditing & Evaluat Lab Islamabad 44000 Pakistan Yeungnam Univ Dept Informat & Commun Engn Gyongsan 38541 South Korea

The predominant Android operating system has captured enormous attention globally not only in smart phone industry but also for varied smart devices. The open architecture and application programming interfaces (APIs) while hosting third party applications has led to explosive growth of varied pervasive sophisticated Android malware production. In this study, we propose a robust, scalable and efficient Cuda-empowered multi-class malware detection technique leveraging Gated Recurrent Unit (GRU) to identify sophisticated Android malware. Experimentation of the proposed technique has been carried out using current state-of-the-art datasets of Android applications (i.e., Android Malware Dataset (AMD), Androzoo). Moreover, to rigorously evaluate the performance of the proposed technique, we have employed standard performance evaluation metrics (e.g., accuracy, precision, recall, F1-score etc.) and compared it with our constructed DL-driven architectures and benchmark algorithms. The GRU-based malware detection system outperforms with 98.99% detection accuracy for malware identification with a trivial trade off in speed efficiency.

关键词： Malware Androids Humanoid robots Smart phones Logic gates Neural networks Feature extraction Android malware deep learning recurrent neural network convolutional neural network deep neural network mobile security

来源：评论

学校读者我要写书评

暂无评论

Deep Neural Architecture Search with Deep Graph Bayesian Optimization 19

Deep Neural Architecture Search with Deep Graph Bayesian Opt...

引用

19th ieee/WIC/ACM International conference on Web Intelligence (WI)

作者： Ma, Lizheng Cui, Jiaxu Yang, Bo Jilin Univ Changchun Jilin Peoples R China

ISBN: (纸本)9781450369343

Image recognition aims to identify objects, places, people, or other targeted items in a given image, and has a wide range of social applications such as natural disasters recognition, plant disease detection, and traffic jam detection. Currently state-of-the-art methods of image recognition are based on deep learning and remain a common pattern in designing and using convolutional neural networks (CNNs). However, designing CNNs is extremely time intensive and requires an expert. Neural architecture search (NAS) can solve this problem by automatically identifying architectures of CNNs that are superior to hand-designed ones. Recently BO has been applied to neural architecture search and shows better performance than pure evolutionary strategies. All these methods adopt Gaussian processes (GPs) as surrogate function, with the handcraft similarity metrics as input. In this work, we propose a Bayesian graph neural network as a new surrogate, which can automatically extract features from deep neural architectures, and use such learned features to fit and characterize black-box objectives and their uncertainty. Based on the new surrogate, we then develop a graph Bayesian optimization framework to address the challenging task of deep neural architecture search. Experiment results show our method significantly outperforms the comparative methods on benchmark tasks.

关键词： deep learning neural architecture search bayesian optimization graph neural network image recognition

来源：评论

学校读者我要写书评

暂无评论

network of ECUs Software Update in Future vehicles 25

Network of ECUs Software Update in Future vehicles

引用

25th ieee International conference on Automation and Computing (ieee ICAC)

作者： Sarwar, Muhammad Hamza Shah, Munam Ali Umair, Muhammad Faraz, Syad Hassnain COMSATS Univ Islamabad Dept Comp Sci Islamabad Pakistan

ISBN: (纸本)9781861376657

Nowadays, the modern vehicles use numerous Electric Control Units (ECUs) which have a complex software architecture. Updating the software in a given ECU is a major challenge since these connected vehicles are prone to security vulnerabilities and cyber-attacks. Another problem is that updating an ECU is a time consuming process and the owners of these vehicles cannot use important features of their cars during this update process. Different software architectures for the ECUs are proposed but none of the existing techniques allow the vehicle driver to use the car during the update. In this paper, we propose an enhanced architecture for the ECUs. The proposed architecture makes use of a delta file which is the actual update in the software. This delta file is placed in the extended RAM and the modules of the software which needs to be updated are run from the delta file. After successful executing of the module from the delta file, the module is updated in the existing software. The proposed ECU architecture allows the owners to use important features of the vehicles during the update process. We run simulation to check the reliability in terms of MTTF due to our non-repairable application of our proposed architecture.

关键词： Software updating in ECU in vehicle ECU network ECU re-programming

来源：评论

学校读者我要写书评

暂无评论

Neural network Architecture Editor With Code Generation

Neural Network Architecture Editor With Code Generation

引用

International Scientific-Practical conference Problems of Infocommunications Science and Technology

作者： Maksym Bekuzarov Mariya Shirokopetleva Oleksandr Samantsov Oksana Mazurova Kharkov National University of Radio Electronics Kharkov Ukraine

This paper discusses the particular properties of modern architectures of neural networks, that lead to complication in their design and implementation. A neural network is a machine learning algorithm, that can be represented as a fixed computational graph with learnable weights on its edges. Modern neural networks have 2 main distinct features - their computation graphs are non-linear and they often consist of repeated blocks. Representing these structures in a programming language is difficult and prone to errors, so a solution to simplify this process was proposed - to develop a graphical editor, which would allow to create, edit, share and generate code for a given Deep Learning framework. An interactive visual representation of the network's architecture as a 2-D graph should greatly simplify the process of designing a neural network. A description of the developed software, its features and internal implementation is given, a comparison with alternative solutions is provided as well.

关键词： Deep learning Visualization Neural networks Computer architecture Tools Software Libraries

来源：评论

学校读者我要写书评

暂无评论

Devise Rust Compiler Optimizations on RISC-V architectures with SIMD Instructions 19

Devise Rust Compiler Optimizations on RISC-V Architectures w...

引用

48th International conference on Parallel Processing (ICPP)

作者： Lin, Heng Chen, Piyo Hwang, Yuan-Shin Lee, Jenq-Kuen Natl Tsing Hua Univ Dept Comp Sci Hsinchu Taiwan Natl Taiwan Univ Sci & Technol Dept Comp Sci & Informat Engn Taipei Taiwan

ISBN: (纸本)9781450371964

Recently, Rust has become a popular system programming language and been widely used in microkernel OS designs, cryptocurrency designs, deep learning applications, and web browsers. Rust is designed for highly safe and concurrent systems and provides similar performance to C++. In the architecture work, RISC-V is with an open instruction set architecture in growing use due to its performance, power efficiency, and open architectures. Extension of ISAs in RISC-V is another advantage. Venders can select the extension ISAs to make the design more flexible. There is SIMD extension ISAs called "V" extension in RISC-V. However, currently Rust was not well-supported on RISC-V platforms. In this paper, we describe how to enable the Rust flow and framework based on LLVM infrastructure for RISC-V architectures with SIMD V extensions. For the basic infrastructure support, we enable two flows for RISC-V architecture. In addition, we also enable RISC-V SIMD instructions for Rust SIMD vector with LLVM support. In the experiment, we illustrate our code generation for Blas style matrix and vector computations for Rust on RISC-V V instructions. The experiment is run on the spike simulator which we implement the V extension for RISC-V architectures.

关键词： Rust RISC-V SIMD LLVM bare metal

来源：评论

学校读者我要写书评

暂无评论

Evaluating Link Latency in Distributed SDN-Based Control Plane architectures

Evaluating Link Latency in Distributed SDN-Based Control Pla...

引用

ieee International conference on Communications (ieee ICC)

作者： Sgambelluri, Andrea Moreno, Xavier Casas Spadaro, Salvatore Monti, Paolo Scuola Super Sant Anna Pisa Italy Tech Univ Catalonia Sch Telecommun Engn Barcelona Spain KTH Royal Inst Technol Sch Elect Engn & Comp Sci Stockholm Sweden Chalmers Inst Technol Dept Elect Engn Gothenburg Sweden

ISBN: (纸本)9781538680889

Being able to account for the value of latency introduced by the transport infrastructure is considered a crucial feature for any 5G service provisioning strategy. This paper proposes a methodology for the automatic monitoring of the link latency in networks with a distributed SDN-based control plane architecture. The proposed link monitoring solution does not introduce any overhead in the openFlow channel and in the data plane. Experimental tests demonstrate the accuracy of the proposed methodology and its ability to scale to large network instances.

关键词： SDN Distributed Control Plane latency open Flow 5G

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Algorithm for Fog Computing Data Processing Decision Optimization

A Dynamic Algorithm for Fog Computing Data Processing Decisi...

引用

ieee International conference on Communications Workshops, ICC

作者： Mohamed Abu Sharkh Mohamad Kalil Digital Media Software Engineering Ferris State University Grand Rapids USA IBM Analytics IBM Canada Ottawa Canada

ISBN: (数字)9781728174402

ISBN: (纸本)9781728174419

Continuing Cloud service challenges like vendor lock-in and security in addition the Cloud clients' increased reliance on the service are motivating large Cloud tenants to build their own Clouds. Cloud providers still have the advantage (cost, availability and expertise) and therefore, a hybrid approach is still the first choice for Cloud clients aspiring to achieve partial independence. This promotes utilizing Fog computing architectures where resources rented from Cloud providers are supplemented by nodes on the edge of network that have limited computational capacity but are closer to request source. Fog nodes offer an opportunity to filter and process requests on the edge of the network in order to distribute the load and minimize the network congestion caused by low-value data. A question is, in turn, posed to Cloud architects on how to optimize the decision as to where to perform each step of the request service. The aim is to mitigate both the risk of pushing high loads to the Cloud servers which increases the cost; and the risk of localizing the whole process and losing the benefits from Cloud services. We build upon our previous work to solve this problem by proposing a novel dynamic programming algorithm to optimize the data processing decision for requests in a Fog environment. The dynamic programming algorithm achieves closer results to the optimal solution by reaching optimal increments of solutions from previous steps in feasible time. Initial experimental results comparing 5 heuristic algorithms are presented with the purpose of offering insight into contradicting factors impacting the problem (cost, network capacity, node and Cloud capacity).

关键词： Cloud computing Delays Heuristic algorithms Optimization Edge computing Data processing Dynamic programming

来源：评论

学校读者我要写书评

暂无评论

Simulation Framework for Virtual Robot programming in Reconfigurable Production Systems 7

Simulation Framework for Virtual Robot Programming in Reconf...

引用

7th CIRP Global Web conference on Towards Shifted Production Value Stream Patterns through Inference of Data, Models, and Technology (CIRPe)

作者： Brecher, Christian Wein, Stephan Xu, Xiaomei Storms, Simon Herfs, Werner Rhein Westfal TH Aachen Werkzeugmaschinenlab WZL Steinbachstr 19 D-52074 Aachen Germany

Flexible, autonomous production systems form the basis for mastering increasing product diversity and customer-specific production. Industrial robots and modularization play a decisive role here, as they support the automation of complex processes and the reconfiguration of production systems. Yet, a change in the mechanical configuration of a production system always results in big efforts for robot programming. Standard teach-in solutions for robot programming are not suitable for highly varying configurations. An automated, adaptive path planning via simulation is necessary to fill this gap and even allow robot programming for non-expert persons. This paper presents a modular framework for automated robot path simulation and virtual robot programming in reconfigurable production systems, based on service oriented architectures and modularization. The framework is evaluated using a reconfigurable production system for flexible cultivation and differentiation of stem cells, the iCellFactory. Within the system, a gantry-mounted six-axis robot transports fluid filled vessels between multiple, freely positioned modules carrying laboratory devices. (C) 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the 7th CIRP Global Web conference

关键词： autonomous production systems robot programming simulation service oriented architecture framework modular reconfigurable 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：