检索结果-内蒙古大学图书馆

arXiv 2017年

作者： Yu, Zhou Yu, Jun Fan, Jianping Tao, Dacheng Key Laboratory of Complex Systems Modeling and Simulation School of Computer Science and Technology Hangzhou Dianzi University China Department of Computer Science University of North Carolina Charlotte United States UBTECH Sydney AI Centre School of IT FEIT University of Sydney Australia

Visual question answering (VQA) is challenging because it requires a simultaneous understanding of both the visual content of images and the textual content of questions. The approaches used to represent the images and questions in a fine-grained manner and questions and to fuse these multimodal features play key roles in performance. Bilinear pooling based models have been shown to outperform traditional linear models for VQA, but their high-dimensional representations and high computational complexity may seriously limit their applicability in practice. For multimodal feature fusion, here we develop a Multi-modal Factorized Bilinear (MFB) pooling approach to efficiently and effectively combine multi-modal features, which results in superior performance for VQA compared with other bilinear pooling approaches. For fine-grained image and question representation, we develop a 'co-attention' mechanism using an end-to-end deep network architecture to jointly learn both the image and question attentions. Combining the proposed MFB approach with co-attention learning in a new network architecture provides a unified model for VQA. Our experimental results demonstrate that the single MFB with co-attention model achieves new state-of-theart performance on the real-world VQA dataset. Code available at https://***/yuzcccc/mfb. Copyright © 2017, The Authors. All rights reserved.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

The Hyper-Kamiokande experiment - snowmass LOI

arXiv

引用

arXiv 2020年

作者： Abe, K. Adrich, P. Aihara, H. Akutsu, R. Alekseev, I. Ali, A. Ameli, F. Anthony, L.H.V. Araya, A. Asaoka, Y. Aushev, V. Bandac, I. Barbi, M. Barr, G. Batkiewicz-Kwasniak, M. Bellato, M. Berardi, V. Bernard, L. Bernardini, E. Berns, L. Bhadra, S. Bian, J. Blanchet, A. Blondel, A. Boiano, A. Bolognesi, S. Bonavera, L. Borjabad, S. Boschi, T. Bose, D. Boyd, S.B. Bozza, C. Bravar, A. Bronner, C. Bubak, A. Buchowicz, A. Avanzini, M. Buizza Cafagna, F.S. Calabria, N.F. Calvo-Mozota, J.M. Cao, S. Catanesi, M.G. Chakraborty, S. Choi, J.H. Choubey, S. Cicerchia, M. Coleman, J. Collazuol, G. Cuen-Rochin, S. Danilov, M. Fuente, E. De la de Perio, P. de Rosa, G. Dealtry, T. Densham, C.J. Dergacheva, A. Deshmukh, N. Devi, M.M. Di Lodovico, F. Di Meo, P. Di Palma, I. Doyle, T.A. Drakopoulou, E. Drapier, O. Dumarchez, J. Eklund, L. El Hedri, S. Ellis, J. Emery, S. Esmaili, A. Fedotov, S. Feng, J. Fernández-Martinez, E. Ferrario, P. Ferrazzi, B. Finch, A. Finley, C. Fiorillo, G. Fitton, M. Friend, M. Fujii, Y. Fukuda, Y. Galinski, G. Gao, J. Garde, C. Garfagnini, A. Garode, S. Gialanella, L. Giganti, C. Gomez-Cadenas, J.J. Gonin, M. González-Nuevo, J. Gorin, A. Gornea, R. Gramegna, F. Grassi, M. Grella, G. Guigue, M. Hadley, D.R. Harada, M. Hartz, M. Hassani, S. Hastings, N.C. Hayato, Y. Hiraide, K. Hoshina, K. Hultqvist, K. Iacob, F. Ichikawa, A.K. Ibnsalih, W. Idrissi Ikeda, M. Inomoto, M. Ioannisian, A. Ishida, T. Ishidoshiro, K. Ishino, H. Ishitsuka, M. Ito, H. Ito, S. Itow, Y. Iwamoto, K. Izumi, N. Izumiyama, S. Jakkapu, M. Jamieson, B. Jang, J.S. Jo, H.S. Jonsson, P. Joo, K.K. Kajita, T. Kakuno, H. Kameda, J. Kano, Y. Karlen, D. Kataoka, Y. Kato, A. Katori, T. Kazarian, N. Khabibullin, M. Khotjantsev, A. Kikawa, T. Kim, J.Y. Kim, S.B. King, S. Kinoshita, T. Kisiel, J. Klekotko, A. Kobayashi, T. Koerich, L. Kolev, N. Konaka, A. Kormos, L.L. Koshio, Y. Kotsar, Y. Kouzakov, K.A. Kowalik, K.L. AGH University of Science and Technology Faculty of Computer Science Electronics and Telecommunications Krakow Poland British Columbia Institute of Technology Physics Department BurnabyBC Canada Università della Campania"L. Vanvitelli" INFN Sezione di Napoli Napoli Italy Carleton University Department of Physics OttawaON Canada University of California Irvine Department of Physics and Astronomy IrvineCA United States Charles University IPNP FMF Prague Czech Republic Chonnam National University Department of Physics Gwangju Korea Republic of Dongshin University Laboratory for High Energy Physics Naju Korea Republic of Donostia International Physics Center Ikerbasque Foundation Basque Country Spain Ecole Polytechnique IN2P3-CNRS Laboratoire Leprince-Ringuet Palaiseau France University of Edinburgh School of Physics and Astronomy Edinburgh United Kingdom ETH Zurich Institute for Particle and Astroparticle Physics Zurich Switzerland University of Geneva Section de Physique DPNC Geneva Switzerland GIST College Gwangju Institute of Science and Technology Gwangju Korea Republic of University of Glasgow School of Physics and Astronomy Glasgow United Kingdom Universidad de Guadalajara CUCEA IT.Ph.D. program GuadalajaraJal Mexico Universidad de Guadalajara CUCEI Departamento de Fisica GuadalajaraJal Mexico H. Niewodniczá nski Institute of Nuclear Physics PAN Cracow Poland Imperial College London Department of Physics London United Kingdom Indian Institute of Technology Guwahati Guwahati India Indian Institute of Technology Jodhpur Department of Physics Karwar Rajasthan India Indian Institute of Technology Kharagpur Department of Physics Kharagpur India INFN Gruppo Collegato di Salerno Fisciano Italy INFN Laboratori Nazionali di Legnaro LegnaroPD Italy INFN Sezione di Bari Università e Politecnico di Bari Bari Italy INFN Sezione di Napoli Università Federico II di Napoli Dipartimento di Fisica Napoli Italy INFN Sezione di Napoli Nap

Hyper-Kamiokande is the next generation underground water Cherenkov detector that builds on the highly successful Super-Kamiokande experiment. The detector which has an 8.4 times larger effective volume than its predecessor will be located along the T2K neutrino beamline and utilize an upgraded J-PARC beam with 2.6 times beam power. Hyper-K’s low energy threshold combined with the very large fiducial volume make the detector unique, that is expected to acquire an unprecedented exposure of 3.8 Mton·year over a period of 20 years of operation. Hyper-Kamiokande combines an extremely diverse science program including nucleon decays, long-baseline neutrino oscillations, atmospheric neutrinos, and neutrinos from astrophysical origins. The scientific scope of this program is highly complementary to liquid-argon detectors for example in sensitivity to nucleon decay channels or supernova detection modes. Hyper-Kamiokande construction has started in early 2020 and the experiment is expected to start operations in 2027. The Hyper-Kamiokande collaboration is presently being formed amongst groups from 19 countries including the United States, whose community has a long history of making significant contributions to the neutrino physics program in Japan. US physicists have played leading roles in the Kamiokande, Super-Kamiokande, EGADS, K2K, and T2K programs. Copyright © 2020, The Authors. All rights reserved.

关键词： Atomic physics

来源：评论

学校读者我要写书评

暂无评论

Mapping urban green open space in Bontang city using QGIS and cloud computing

引用

IOP Conference Series: Earth and Environmental science 2018年第1期144卷

作者： F Agus Ramadiani W Silalahi A Armanda Kusnandar GIS and Environment Modeling Research Group Computer Science Department Faculty of Computer Science and Information Technology-CSIT Mulawarman University Samarinda Kalimantan Timur Province Indonesia Informatics Department STMIK Widya Cipta Dharma Samarinda Kalimantan Timur Province Indonesia

Digital mapping techniques are available freely and openly so that map-based application development is easier, faster and cheaper. A rapid development of Cloud Computing Geographic Information System makes this system can help the needs of the community for the provision of geospatial information online. The presence of urban Green Open Space (GOS) provide great benefits as an oxygen supplier, carbon-binding agent and can contribute to providing comfort and beauty of city life. This study aims to propose a platform application of GIS Cloud Computing (CC) of Bontang City GOS mapping. The GIS-CC platform uses the basic map available that's free and open source. The research used survey method to collect GOS data obtained from Bontang City Government, while application developing works Quantum GIS-CC. The result section describes the existence of GOS Bontang City and the design of GOS mapping application.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Hoeffding Tree Algorithms for Anomaly Detection in Streaming Datasets: A Survey

引用

Journal of Information Security 2017年第4期8卷 339-361页

作者： Asmah Muallem Sachin Shetty Jan Wei Pan Juan Zhao Biswajit Biswal Department of Electrical and Computer Engineering Tennessee State University Nashville TN USA Virginia Modeling Analysis and Simulation Center Old Dominion University Norfolk VA USA R&D Engineering Department Roadstar.AI. Mountain View CA USA Department of Electrical and Computer Engineering Tennessee State University Nashville TN USA Department of Mathematics and Computer Science South Carolina State University Orangeburg SC USA

This survey aims to deliver an extensive and well-constructed overview of using machine learning for the problem of detecting anomalies in streaming datasets. The objective is to provide the effectiveness of using Hoeffding Trees as a machine learning algorithm solution for the problem of detecting anomalies in streaming cyber datasets. In this survey we categorize the existing research works of Hoeffding Trees which can be feasible for this type of study into the following: surveying distributed Hoeffding Trees, surveying ensembles of Hoeffding Trees and surveying existing techniques using Hoeffding Trees for anomaly detection. These categories are referred to as compositions within this paper and were selected based on their relation to streaming data and the flexibility of their techniques for use within different domains of streaming data. We discuss the relevance of how combining the techniques of the proposed research works within these compositions can be used to address the anomaly detection problem in streaming cyber datasets. The goal is to show how a combination of techniques from different compositions can solve a prominent problem, anomaly detection.

关键词： Hoeffding Trees Distributed Ensembles Anomaly Detection Machine Learning Spark

来源：评论

学校读者我要写书评

暂无评论

Autonomous Navigation for an Unmanned Aerial Vehicle by the Decomposition Coordination Method 4

Autonomous Navigation for an Unmanned Aerial Vehicle by the ...

引用

International Conference on Electrical Engineering, computer science and Informatics

作者： Chaimaa Jihane Hala El Ouarrak Mohamed Mestari Mostafa Rachik Laboratory of Signals Distributed Systems and Artificial Intelligence (SSDIA) Ecole Normale Suprieure d'Enseignement Technique Mohammedia Av Hassan II Mohammedia Morocco Department of Mathematics and Computer Sciences Laboratory of Analysis Modeling and Simulation (LAMS) Faculty of Science Ben M'Sik Av Driss El Harti B.P 7955 Sidi Othmane Casablanca Morocco

ISBN: (纸本)9781538605509;9781538605493

This paper introduces a new approach for solving the navigation problem of Unmanned Aerial Vehicles (UAV) by studying its rotational and translational dynamics and then solving the nonlinear model by the Decomposition Coordination method. The objective is to reach a destination goal by the mean of an autonomous computed optimal path calculated through optimal control sequence. Solving such complex systems often requires a great amount of computation. However, the approach considered herein is based on the Decomposition Coordination principle, which allows the nonlinearity to be treated at a local level, thus offering a low computing time. The stability of the method is discussed with sufficient conditions for convergence. A numerical application is given in consolidation the theoretical results.

关键词： considered herein Decomposition theoretical

来源：评论

学校读者我要写书评

暂无评论

Beyond bilinear: Generalized multimodal factorized high-order pooling for visual question answering

arXiv

引用

arXiv 2017年

作者： Yu, Zhou Yu, Jun Xiang, Chenchao Fan, Jianping Tao, Dacheng Key Laboratory of Complex Systems Modeling and Simulation School of Computer Science and Technology Hangzhou Dianzi University China Department of Computer Science University of North Carolina Charlotte United States UBTECH Sydney Artificial Intelligence Centre School of Information Technologies Faculty of Engineering and Information Technologies University of Sydney DarlingtonNSW2008 Australia

Visual question answering (VQA) is challenging because it requires a simultaneous understanding of both visual content of images and textual content of questions. To support the VQA task, we need to find good solutions for the following three issues: 1) fine-grained feature representations for both the image and the question;2) multi-modal feature fusion that is able to capture the complex interactions between multi-modal features;3) automatic answer prediction that is able to consider the complex correlations between multiple diverse answers for the same question. For fine-grained image and question representations, a 'co-attention' mechanism is developed by using a deep neural network architecture to jointly learn the attentions for both the image and the question, which can allow us to reduce the irrelevant features effectively and obtain more discriminative features for image and question representations. For multi-modal feature fusion, a generalized Multi-modal Factorized High-order pooling approach (MFH) is developed to achieve more effective fusion of multi-modal features by exploiting their correlations sufficiently, which can further result in superior VQA performance as compared with the state-of-the-art approaches. For answer prediction, the KL (Kullback-Leibler) divergence is used as the loss function to achieve precise characterization of the complex correlations between multiple diverse answers with same or similar meaning, which can allow us to achieve faster convergence rate and obtain slightly better accuracy on answer prediction. A deep neural network architecture is designed to integrate all these aforementioned modules into a unified model for achieving superior VQA performance. With an ensemble of our MFH models, we achieve the state-of-the-art performance on the large-scale VQA datasets and win the runner-up in VQA Challenge 2017. Copyright © 2017, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Energy efficiency comparison of hypervisors 7

Energy efficiency comparison of hypervisors

引用

7th International Green and Sustainable Computing Conference, IGSC 2016

作者： Jiang, Congfeng Ou, Dongyang Wang, Yumei You, Xindong Zhang, Jilin Wan, Jian Luo, Bing Shi, Weisong Key Laboratory of Complex Systems Modeling and Simulation Hangzhou Dianzi University Zhejiang China Mobile and Internet Systems Laboratory Department of Computer Science Wayne State University DetroitMI United States

ISBN: (纸本)9781509051175

Current cloud data centers are fully virtualized for service consolidations and power/energy reduction. Although virtualization could reduce real time power and overall energy consumption, the energy characteristics of hypervisors hosting different workloads are not well profiled and understood. In this paper, we investigate the power and energy characteristics of mainstream hypervisors and container engine, i.e., VMware ESXi, Microsoft Hyper-V, KVM, XenServer and Docker, on five different platforms (two mainstream 2U rack servers, one emerging ARM64 server, one desktop server, and one laptop) with hundreds of hours' power measures. We use both computing intensive and mixed web server-database workloads to explore the power and energy characteristics of different hypervisors. Extensive experiment results of four workload levels (very light, light, fair, and very heavy workload) demonstrate that hypervisors expose different power and energy characteristics. We find that: (1) Hypervisors expose different power and energy consumption on the same hardware running same workloads. (2) Although mainstream hypervisors have different energy efficiencies aligned with different workload types and workload levels, there is no single hypervisor that outperforms all other hypervisors on all platforms in terms of power or energy consumptions. (3) Although container virtualization is considered as light-weight virtualization in terms of implementation and maintenance, it is not essentially more power efficient than traditional virtualization technology. (4) ARM64 server does have lower power consumption, but they finish computing jobs with longer execution time and sometimes consume more energy. And ARM64 servers has medium energy consumption per database operations for mixed workloads. The results presented in this paper provide useful insights to system designers, as well as data center operators for power-aware workload placement and virtual machine scheduling. © 2016 IEEE.

关键词： Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

Building a hybrid DEVS and GRASS model using a composable cellular automaton

引用

International Journal of modeling, simulation, and Scientific Computing 2016年第1期7卷 62-92页

作者： Gary R.Mayer Hessam S.Sarjoughian Department of Computer Science Southern Illinois University Edwardsville 61 Circle Drive EdwardsvilleIllinois 62026USA Arizona Center for Integrative Modeling&Simulation School of Computing Informaticsand Decision Systems Engineering Arizona State University699 South Mill Avenue TempeArizona 85281USA

modeling and simulation is pervasive throughout many different *** computing technology has provided more capability,the systems being modeled and simulated have grown larger and more *** times,these large systems are managed as interacting *** it is necessary for the simulation to allow disparate subsystems to maintain their independence,then a hybrid model of the subsystems should be ***,to ease the burden of verification and validation of simulation results,a proven system theoretical modeling specification should be ***,many communities have already adopted nonsystem theoretical software solutions and established a group of domain experts familiar with these *** paper provides two things:a formal approach to building a hybrid model,and a discussion of how to incorporate a nonsystem theoretical software implementation into a proven *** first is done through the implementation of a Knowledge Interchange Broker(KIB)as an Interaction Model(IM).The second is accomplished by exemplifying the use of the IM in an agent-environment hybrid *** the hybrid model,the agent is implemented in the Discrete-event System(DEVS)specification and the environment is implemented in the Geographical Resources Analysis Support System(GRASS)using a Composable Cellular Automaton(CCA)*** concept has been successfully applied to both example models and an interdisciplinary research project where the interactions between human activities and landscape processes are studied.

关键词： Multi-formalism modeling hybrid simulation composability agent cellular automata

来源：评论

学校读者我要写书评

暂无评论

Demonstrating Open science for modeling and simulation Research 20

Demonstrating Open Science for Modeling and Simulation Resea...

引用

20th IEEE/ACM International Symposium on Distributed simulation and Real Time Applications, DS-RT 2016

作者： Taylor, Simon J. E. Fabiyi, Adedeji Anagnostou, Anastasia Barbera, Roberto Torrisi, Mario Ricceri, Rita Becker, Bruce Modelling and Simulation Group Department of Computer Science Brunel University London United Kingdom Department of Physics and Astronomy University of Catania Italy Council for Scientific and Industrial Research Capetown South Africa

ISBN: (纸本)9781509035045

Most conference and journal publications present contributions based on research artefacts (data, results, software, etc.) that are difficult for researchers to access. Open science aims to promote open access to research presented in academic works. Ideally, the software, data and results presented in a scientific article should be available for other scientists to use, validate and build upon for their own research. This is particularly true in some modeling and simulation (M and S) research where in addition to the above access might also be required to complex models. Using a case study based on an Infection Model, this demonstration shows how Open science approaches based on Digital Object Identifiers, Researcher Registries, Open Access Data Repositories, Scientific Gateways and e-Infrastructures project can support M and S research. © 2016 IEEE.

关键词： e-Infrastructure e-Infrastructure modeling Open Access Open Data Open science Research simulation

来源：评论

学校读者我要写书评

暂无评论

Mixed optimization for constrained resource allocation, an application to a local bus service 16

Mixed optimization for constrained resource allocation, an a...

引用

simulation Winter Conference

作者： Felisa J. Vázquez-Abad Larry Fenn Department of Computer Science CUNY Institute for Computer Simulation Stochastic Modeling and Optimization New York NY USA

ISBN: (纸本)9781509044849

The present paper follows up on Vázquez-Abad (2013), where we applied the ghost simulation model to a public transportation problem. The ghost simulation model replaces faster point processes (passenger arrivals) with a “fluid” model while retaining a discrete event simulation for the rest of the processes (bus dynamics). This is not an approximation, but an exact conditional expectation when the fast process is Poisson. It can be interpreted as a Filtered Monte Carlo method for fast simulation. In the current paper we develop the required theory to implement a mixed optimization procedure to find the optimal fleet size under a stationary probability constraint. It is a hybrid optimization because for each fleet size, the optimal headway is real-valued, while the fleet size is integer-valued. We exploit the structure of the problem to implement a stopped target tracking method combined with stochastic binary search.

关键词： Optimization Stochastic processes Computational modeling Resource management Search problems Investment Public transportation

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：