检索结果-内蒙古大学图书馆

arXiv 2017年

作者： Akopyan, Manuk Khashba, Eshsou Institute for System Programming

Video classification problem has been studied many years. The success of Convolutional Neural Networks (CNN) in image recognition tasks gives a powerful incentive for researchers to create more advanced video classification approaches. As video has a temporal content Long Short Term Memory (LSTM) networks become handy tool allowing to model long-term temporal clues. Both approaches need a large dataset of input data. In this paper three models provided to address video classification using recently announced YouTube-8M large-scale dataset. The first model is based on frame pooling approach. Two other models based on LSTM networks. Mixture of Experts intermediate layer is used in third model allowing to increase model capacity without dramatically increasing computations. The set of experiments for handling imbalanced training data has been conducted. Copyright © 2017, The Authors. All rights reserved.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Test Program Generator MicroTESK for RISC-V

Test Program Generator MicroTESK for RISC-V

引用

International Workshop on Microprocessor Test and Verification (MTV)

作者： Mikhail Chupilko Alexander Kamkin Artem Kotsynyak Alexander Protsenko Sergey Smolov Andrei Tatarnikov Ivannikov Institute for System Programming of the Russian Academy of Sciences (ISP RAS) Lomonosov Moscow State University (MSU) National Research University - Higher School of Economics (HSE)

ISBN: (纸本)9781538692516

The paper presents a test program generator for functional verification of RISC-V microprocessors. The generator is implemented on the base of MicroTESK framework and consists of formal specifications of RISC-V ISA and ISA-independent core. The specifications describe instructions' syntax and semantics and can be easily modified to support more instructions (including custom extensions). The core implements techniques of instruction sequences composition and test data generation. Test programs are generated from test templates, describing the programs' structural and behavioral properties; among generation techniques, random, combinatorial, and constraint-based ones are supported.

关键词： Registers Microprocessors Object oriented modeling Generators Syntactics Standards Computer architecture

来源：评论

学校读者我要写书评

暂无评论

Preface

Lecture Notes in Computer Science (including subseries Lectu...

引用

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 2019年 11771 LNCS卷 v页

作者： Mazzara, Manuel Bruel, Jean-Michel Meyer, Bertrand Petrenko, Alexander Innopolis University Innopolis Russia IUT de Blagnac Blagnac France Ivannikov Institute for System Programming Russian Academy of Sciences Moscow Russia

来源：评论

学校读者我要写书评

暂无评论

Secure Verifiable Secret Short Sharing Scheme for Multi-Cloud Storage

Secure Verifiable Secret Short Sharing Scheme for Multi-Clou...

引用

International Conference on High Performance Computing and Simulation

作者： Maxim Deryabin Nikolay Chervyakov Andrei Tchernykh Mikhail Babenko Nikolay Kucherov Vanessa Miranda-Lopez Arutyun Avetisyan North-Caucasus Federal University CICESE Research Center Ensenada Mexico South Ural State University CICESE Research Center Institute for System Programming of the Russian Academy of Sciences

ISBN: (纸本)9781538678800

In this paper, we propose a new approach to the construction of computationally secure secret sharing scheme based on the simultaneous use of the Residue Number system (RNS), as the symmetric encryption, and the perfect Asmuth-Bloom secret sharing scheme. We combine the useful features of Redundant RNS to design space efficient secret sharing scheme with enough level of security and to control of data integrity. This combination provides verifiability and achieves a high speed of data processing.

关键词： Threshold secret sharing scheme Computationally secure secret sharing Verifiable scheme Residue Number system Cloud data storage Distributed storage Multi-clouds

来源：评论

学校读者我要写书评

暂无评论

Lemma functions for frama-C: C programs as proofs

arXiv

引用

arXiv 2018年

作者： Volkov, Grigoriy Mandrykin, Mikhail Efremov, Denis Faculty of Computer Science National Research University Higher School of Economics Moscow Russia Software Engineering Department Ivannikov Institute for System Programming Russian Academy of Sciences Moscow Russia

This paper describes the development of an auto-active verification technique in the Frama-C framework. We outline the lemma functions method and present the corresponding ACSL extension, its implementation in Frama-C, and evaluation on a set of string-manipulating functions from the Linux kernel. We illustrate the benefits our approach can bring concerning the effort required to prove lemmas, compared to the approach based on interactive provers such as Coq. Current limitations of the method and its implementation are discussed. Copyright © 2018, The Authors. All rights reserved.

关键词： Formal verification

来源：评论

学校读者我要写书评

暂无评论

Deductive verification of unmodified linux kernel library functions

arXiv

引用

arXiv 2018年

作者： Efremov, Denis Mandrykin, Mikhail Khoroshilov, Alexey National Research University Higher School of Economics Moscow Russia Ivannikov Institute for System Programming RAS Moscow Russia Moscow Institute of Physics and Technology Moscow Russia Lomonosov Moscow State University Moscow Russia

This paper presents results from the development and evaluation of a deductive verification benchmark consisting of 26 unmodified Linux kernel library functions implementing conventional memory and string operations. The formal contract of the functions was extracted from their source code and was represented in the form of preconditions and postconditions. The correctness of 23 functions was completely proved using AstraVer toolset, although success for 11 functions was achieved using 2 new specification language constructs. Another 2 functions were proved after a minor modification of their source code, while the final one cannot be completely proved using the existing memory model. The benchmark can be used for the testing and evaluation of deductive verification tools and as a starting point for verifying other parts of the Linux kernel. Copyright © 2018, The Authors. All rights reserved.

关键词： Formal verification

来源：评论

学校读者我要写书评

暂无评论

Numerical simulation of jet flows in homogeneous and heterogeneous media

引用

Journal of Physics: Conference Series 2019年第1期1404卷

作者： A K Shevchenko S V Strijhak S N Yakovenko Khristianovich Institute of Theoretical and Applied Mechanics Siberian Branch of Russian Academy of Sciences Novosibirsk 630090 Russia Ivannikov Institute for System Programming Russian Academy of Sciences Moscow 109004 Russia

Submerged jets at different Reynolds numbers (Re) with random perturbations added to the inlet velocity profile are simulated numerically by solving the Navier–Stokes equations. The effects of Re on flow behavior and the length of laminar jet region are studied. A natural gas jet flow emitting form the sea bottom is also simulated using the URANS and VOF methods, and forces acting on an object placed within the multiphase jet are evaluated.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Monitoring and accounting for the distributed computing system of the ATLAS experiment 27

Monitoring and accounting for the distributed computing syst...

引用

27th Symposium on Nuclear Electronics and Computing, NEC 2019

作者： Barberis, D. Aimar, A. Alekseev, A. de Sousa Andrade, P.M. Rodrigues Beermann, T.A. Gardner, R.W. Bear, B. Garrido Korchuganova, T. Magnoni, L. Padolski, S. Schanet, E. Tsvetkov, N. Vukotić, I. Wenaus, T. Dipartimento di Fisica dell’Università di Genova e INFN Sezione di Genova Via Dodecaneso 33 GenovaI - 16146 Italy CERN Genève 23CH - 1211 Switzerland Institute of System Programming Russian Academy of Science Moscow Russia Universidad Andrés Bello Santiago Chile Plekhanov Russian University of Economics Stremyanny Lane 36 MoscowRU – 117997 Russia Bergische Universitaet Wuppertal Gaußstr. 20 WuppertalDE - 42119 Germany University of Chicago Enrico Fermi Institute 5640 S Ellis Ave ChicagoIL60637 United States Brookhaven National Laboratory UptonNY United States Fakultät für Physik Ludwig-Maximilians-Universität München Am Coulombwall 1 Garching bei MünchenDE - 85748 Germany

Over the years, ATLAS has developed a large number of monitoring and accounting tools for distributed computing applications. In advance of the increased experiment data rates and monitoring data volumes foreseen for LHC Run 3, starting in 2012, a new infrastructure has been provided by the CERN-IT Monit group, based on InfluxDB as the data store and Grafana as the display environment. ATLAS is adapting and further developing its monitoring tools to use this infrastructure for data and workflow management monitoring and accounting dashboards, expanding the range of previous possibilities with the aim of achieving a single, simpler, environment for all monitoring applications. This contribution describes the tools used, the data flows for monitoring and accounting, the problems encountered and the solutions found. Copyright © 2019 for this paper by its authors.

关键词： Monitoring

来源：评论

学校读者我要写书评

暂无评论

Multi-Agent Approach for Dynamic Elasticity of Virtual Machines Provisioning in Heterogeneous Distributed Computing Environment

Multi-Agent Approach for Dynamic Elasticity of Virtual Machi...

引用

International Conference on High Performance Computing & Simulation (HPCS)

作者： Alexander Feoktistov Ivan Sidorov Andrei Tchernykh Alexei Edelev Valery Zorkalzev Roman Kostromin Sergey Gorsky Igor Bychkov Arutyun Avetisyan Matrosov Institute for System Dynamics and Control Theory of the Siberian Branch of the Russian Academy of Sciences Irkutsk Russia Matrosov Inst. for Syst. Dynamics & Control Theor. Irkutsk Russia Ivannikov Institute for System Programming of the Russian Academy of Sciences Moscow Russia Melentiev Energy Systems Institute of the Siberian Branch of the Russian Academy of Sciences Irkutsk Russia

ISBN: (纸本)9781538678800

We address dynamic elasticity issues of VM provisioning in a heterogeneous distributed computing environment that integrates resources of a data center. We consider the scenario when the center includes dedicated resources (private cloud) for providing virtualized service and non-dedicated resources for Grid-computing. Existing platforms for the resource virtualization does not support management of such an infrastructure. We propose an approach to a job management based on the dynamic elasticity of virtual machines provisioning using resources of both types. We develop the multi¬agent job scheduler for dedicated resources and hypervisor shell to launch virtual machines through queues of resource management systems in non-dedicated resources. The scheduler provides a dynamic elasticity of virtual machine provisioning. Advantages of the offered approach to the resource virtualization are demonstrated by an example of a job flow management for a scalable application to solve the complicated practical problem. It is related to the energy security of Vietnam. Provided experiments show that using the developed tools together with the platform for the resource virtualization enables agents to significantly speed up the problem-solving process.

关键词： Cloud computing Resource management Dynamic scheduling Virtual machine monitors Elasticity Virtual machining

来源：评论

学校读者我要写书评

暂无评论

Transfer Learning for Morphological Tagging in Russian

Transfer Learning for Morphological Tagging in Russian

引用

Ivannikov ISPRAS Open Conference (ISPRAS)

作者： Ivan Andrianov Vladimir Mayorov Ivannikov Institute for System Programming of the Russian Academy of Sciences

This paper is devoted to morphological tagging task for Russian. There are multiple corpora created for the task during the years of research. Unfortunately, these corpora often have incompatible annotation guidelines and tag sets. This makes it difficult to utilize more than one corpus for training machine learning methods. Several attempts to manually unify guidelines and tag sets took place in the past. The most recent one was MorphoRuEval-2017 evaluation where organizers decided to unify annotations of four corpora. In this paper we propose a morphological tagging method which is able to be trained on any set of corpora with arbitrary guidelines and tag sets without any manual mapping provided. The method is based on transfer learning technique successfully applied in multilingual and cross-domain research. Firstly we establish a couple of solid neural network-based morphological tagging baselines. They both employ bidirectional LSTM networks but utilize different word embeddings. Comparative study on multiple corpora shows that fasttext model clearly outperforms word2vec one. Further we extend these baselines with transfer learning application. We prove transfer learning-based method effectiveness by performing two series of experiments. In the first series we exploit three wellknown corpora with incompatible tag sets: RNC, SynTagRus (original and Universal Dependencies versions). In the second series we utilize four corpora from MorphoRuEval-2017 evaluation with tag sets preliminary unified by organizers. In both cases transfer learning application improves performance of the method while training on more corpora. In the second case it even surpasses traditional joint training.

关键词： Tagging Task analysis Neural networks Training Guidelines Learning systems Speech

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：