检索结果-内蒙古大学图书馆

arXiv 2018年

作者： Lu, Hang Wei, Xin Lin, Ning Yan, Guihai Li, Xiaowei State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences University of Chinese Academy of Sciences

Inference efficiency is the predominant consideration in designing deep learning accelerators. Previous work mainly focuses on skipping zero values to deal with remarkable ineffectual computation, while zero bits in non-zero values, as another major source of ineffectual computation, is often ignored. The reason lies on the difficulty of extracting essential bits during operating multiply-and-accumulate (MAC) in the processing element. Based on the fact that zero bits occupy as high as 68.9% fraction in the overall weights of modern deep convolutional neural network models, this paper firstly proposes a weight kneading technique that could eliminate ineffectual computation caused by either zero value weights or zero bits in non-zero weights, simultaneously. Besides, a split-and-accumulate (SAC) computing pattern in replacement of conventional MAC, as well as the corresponding hardware accelerator design called Tetris are proposed to support weight kneading at the hardware level. Experimental results prove that Tetris could speed up inference up to 1.50x, and improve power efficiency up to 5.33x compared with the state-of-the-art baselines. Copyright © 2018, 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

Electron-phonon coupling in d-electron solids: A temperature-dependent study of rutile TiO2 by first-principles theory and two-photon photoemission

arXiv

引用

arXiv 2020年

作者： Shang, Honghui Argondizzo, Adam Tan, Shijing Zhao, Jin Rinke, Patrick Carbogno, Christian Scheffler, Matthias Petek, Hrvoje Fritz-Haber-Institut der Max-Planck-Gesellschaft Faradayweg 46 Berlin-DahlemD-14195 Germany Department of Physics and Astronomy Pittsburgh Quantum Institute University of Pittsburgh PittsburghPA15260 United States Department of Physics University of Science and Technology of China Hefei230026 China Department of Applied Physics Aalto University P.O. Box 11100 AaltoFI-00076 Finland State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China

Rutile TiO2 is a paradigmatic transition-metal oxide with applications in optics, electronics, photocatalysis, etc., that are subject to pervasive electron-phonon interaction. To understand how energies of its electronic bands, and in general semiconductors or metals where the frontier orbitals have a strong d-band character, depend on temperature, we perform a comprehensive theoretical and experimental study of the effects of electron-phonon (e-p) interactions. In a two-photon photoemission (2PP) spectroscopy study we observe an unusual temperature dependence of electronic band energies within the conduction band of reduced rutile TiO2, which is contrary to the well-understood sp-band semiconductors and points to a so far unexplained dichotomy in how the e-p interactions affect differently the materials where the frontier orbitals are derived from the sp- and d orbitals. To develop a broadly applicable model, we employ state-of-the-art first-principles calculations that explain how phonons promote interactions between the Ti-3d orbitals of the conduction band within the octahedral crystal field. The characteristic difference in e-p interactions experienced by the Ti-3d orbitals of rutile TiO2 crystal lattice are contrasted with the more familiar behavior of the Si-2s orbitals of stishovite SiO2 polymorph, in which the frontier 2s orbital experiences a similar crystal field with the opposite effect. The findings of this analysis of how e-p interactions affect the d- and sp-orbital derived bands can be generally applied to related materials in a crystal field. The calculated temperature dependence of d-orbital derived band energies agrees well with and explains the temperature-dependent inter-d-band transitions recorded in 2PP spectroscopy of TiO2. The general understanding of how e-pinteractions affect d-orbital derived bands is likely to impact the understanding of temperature-dependent properties of highly correlated materials. Copyright © 2020, The Authors. All r

关键词： Temperature distribution

来源：评论

学校读者我要写书评

暂无评论

A multi-scale mapping approach based on a deep learning CNN model for reconstructing high-resolution urban DEMs

arXiv

引用

arXiv 2019年

作者： Jiang, Ling Hu, Yang Xia, Xilin Liang, Qiuhua Soltoggio, Andrea State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering Hohai University Nanjing210098 China Anhui Engineering Laboratory of Geo-information Smart Sensing and Services Chuzhou University Chuzhou239000 China School of Architecture Building and Civil Engineering Loughborough University Leicestershire United Kingdom School of Computer Science Loughborough University Leicestershire United Kingdom

The shortage of high-resolution urban digital elevation model (DEM) datasets has been a challenge for modelling urban flood and managing its risk. A solution is to develop effective approaches to reconstruct high-resolution DEMs from their low-resolution equivalents that are more widely available. However, the current high-resolution DEM reconstruction approaches mainly focus on natural topography. Few attempts have been made for urban topography which is typically an integration of complex man-made and natural features. This study proposes a novel multi-scale mapping approach based on convolutional neural network (CNN) to deal with the complex characteristics of urban topography and reconstruct high-resolution urban DEMs. The proposed multi-scale CNN model is firstly trained using urban DEMs that contain topographic features at different resolutions, and then used to reconstruct the urban DEM at a specified (high) resolution from a low-resolution equivalent. A two-level accuracy assessment approach is also designed to evaluate the performance of the proposed urban DEM reconstruction method, in terms of numerical accuracy and morphological accuracy. The proposed DEM reconstruction approach is applied to a 121 km2 urbanized area in London, UK. Compared with other commonly used methods, the current CNN based approach produces superior results, providing a cost-effective innovative method to acquire high-resolution DEMs in other data-scarce environments. Copyright © 2019, The Authors. All rights reserved.

关键词： Surveying

来源：评论

学校读者我要写书评

暂无评论

Design and realization of rock salt gas storage database management system based on SQL Server

引用

Petroleum 2018年第4期4卷 466-472页

作者： Yingjie Wang Jianjun Liu Xiang He Bing Wang State Key Laboratory of Oil&Gas Reservoir Geology and Exploitation SouthWest Petroleum UniversityChengduChina School of Geoscience and Technology SouthWest Petroleum UniversityChengduChina School of Civil Engineering and Architecture Wuhan Polytechnic UniversityWuhanChina School of Computer Science SouthWest Petroleum UniversityChengduChina

Projects involving the construction of rock salt underground gas storage have several disadvantages,for example,effective management is not employed to manage the production information and data in the process of the project,resulting in duplication of data storage,waste of storage space,lower efficiency of data calling,and negative effects on the efficiency of data ***,a database and its management systems for a rock salt gas storage was constructed based on an SQL Server database system,primarily including the management forms of the geological modeling,storage simulation,stability evaluation,economic evaluation,and covering the addition and delete checks of static and dynamic *** security of the system was improved by setting the administrator *** establishment of the database management system was of tremendous importance and it provided a significant technical support for the development of the gas storage project.

关键词： computer technology and application Database system SQL Server Rock salt gas storage

来源：评论

学校读者我要写书评

暂无评论

Optimizing network efficiency of dataflow architectures through dynamic packet merging

Optimizing network efficiency of dataflow architectures thro...

引用

International Green Computing Conference

作者： Yujing Feng Han Li Xu Tan Xiaochun Ye Dongrui Fan Zhimin Tang School of Computer and Control Engineering University of Chinese Academy of Sciences Beijing China State Key Laboratory of Computer Architecture Chinese Academy of Sciences Beijing China

ISBN: (纸本)9781538674673

Dataflow processor has shown its unique advantages in executing high performance computing applications with its communication-exposed microarchitecture. In dataflow processors, large amounts of data are directly transferred between instructions through a network-on-chip. The efficiency of data transfer is an imperative performance metric that needs to be optimized in most dataflow processors. Based on the specific features of the dataflow network, we propose a mechanism for dynamically merging the packets in the routers. By testing workloads with varying characteristics, the experiment results demonstrate that the average latency of data transfer is reduced by 11.8%, the performance of dataflow accelerator is improved by 14.0%.

关键词： Merging Routing Data transfer Bandwidth Data models Arrays

来源：评论

学校读者我要写书评

暂无评论

AIBench: An industry standard internet service ai benchmark suite

arXiv

引用

arXiv 2019年

作者： Gao, Wanling Tang, Fei Wang, Lei Zhan, Jianfeng Lan, Chunxin Luo, Chunjie Huang, Yunyou Zheng, Chen Dai, Jiahui Cao, Zheng Zheng, Daoyi Tang, Haoning Zhan, Kunlin Wang, Biao Kong, Defei Wu, Tong Yu, Minghe Tan, Chongkang Li, Huan Tian, Xinhui Li, Yatao Lu, Gang Shao, Junchao Wang, Zhenyu Wang, Xiaoyu Ye, Hainan State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing Academy of Frontier Sciences and Technology University of Chinese Academy of Sciences Alibaba Baidu Tencent *** NetEase ByteDance China National Institute of Metrology Zhihu Lenovo Paypal Moqi Microsoft Research Asia Huawei *** CloudTa Intellifusion

Today's Internet Services are undergoing fundamental changes and shifting to an intelligent computing era where AI is widely employed to augment services. In this context, many innovative AI algorithms, systems, and architectures are proposed, and thus the importance of benchmarking and evaluating them rises. However, modern Internet services adopt a microservice-based architecture and consist of various modules. The diversity of these modules and complexity of execution paths, the massive scale and complex hierarchy of datacenter infrastructure, the confidential issues of data sets and workloads pose great challenges to benchmarking. Copyright © 2019, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Polymorphic PUF: Exploiting reconfigurability of CPU+FPGA SoC to resist modeling attack

Polymorphic PUF: Exploiting reconfigurability of CPU+FPGA So...

引用

2017 Asian Hardware Oriented Security and Trust Symposium, AsianHOST 2017

作者： Ye, Jing Gong, Yue Hu, Yu Li, Xiaowei State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China

ISBN: (纸本)9781538614211

Physical Unclonable Function (PUF) is a promising hardware security primitive. However, it is severely threatened by modeling attacks. This paper proposes a novel Polymorphic PUF for CPU+FPGA SoC. We fully exploit the dynamic reconfigurability of the SoC and use the partial reconfiguration bitstream as the challenge to minimize the Challenge Response Pair (CRP) correlation so as to resist modeling attacks. An asymmetric RO pair is proposed to produce the response. Experiments on real CPU+FPGA SoCs show the high resistance of Polymorphic PUF against modeling attacks, with good uniformity and uniqueness. © 2017 IEEE.

关键词： System-on-chip

来源：评论

学校读者我要写书评

暂无评论

Research on the Construction Method of Domain Semantic Knowledge Base Based on Freebase

Research on the Construction Method of Domain Semantic Knowl...

引用

International Conference on Information Systems and computer Aided Education (ICISCAE)

作者： Deyan Chen Hong Zhao State Key Laboratory of Software Architecture (Neusoft Corporation) Shenyang China School of Computer Science & Engineering Northeastern University Shenyang China

ISBN: (纸本)9781538657393;9781538657386

In the process of constructing a domain semantic knowledge base, reusing existing domain knowledge bases not only facilitates the sharing, integration and reuse of the domain semantic knowledge base, but also can accelerate the construction of the domain semantic knowledge base. The open and fast growing Freebase database is a good data source which can be reused to construct the domain semantic knowledge base. However, extracting domain knowledge from the Freebase RDF dump faces many challenges. Aiming at these challenges and the shortcomings of existing research, this paper proposes a method to extract domain knowledge quickly, accurately and completely from the Freebase RDF dump and describe the domain knowledge using the semantic constructs in ontology standard description languages. Finally, the method of this paper is evaluated, including the comparison and analysis of related methods with work objectives, processes, software tools used, processing results and processing performance.

关键词： Resource description framework Semantics Ontologies Knowledge based systems Standards Loading Data mining

来源：评论

学校读者我要写书评

暂无评论

X-Diag: Automated Debugging Cross-Browser Issues in Web Applications

X-Diag: Automated Debugging Cross-Browser Issues in Web Appl...

引用

IEEE International Conference on Web Services (ICWS)

作者： Shaopeng Xu Chenyu Zhou Zhiwei Gu Guoquan Wu Wei Chen Jun Wei Universitv of Chinese Academv of Sciences Beijing University of Civil Engineering and Architecture State Key Laboratory of Computer Science Chinese Academy of Sciences

With the advent of Web 2.0 application, and the increasing number of browsers and platforms on which the applications can be executed, cross-browser incompatibilities (XBIs) are becoming a serious problem for organizations to develop web-based software. Although many techniques and tools have been proposed to detect cross-browser issues, there still lacks a comprehensive approach to locate the root causes of various cross-browser issues. To address this limitation, this paper proposes X-Diag, an automated technique for debugging XBIs based on our findings from an extensive study of the root causes of XBIs in real-world applications. The characteristic of X-Diag is that it narrows down the root causes of cross-browser issues step-by-step by checking whether such issues are caused by incompatible DOM APIs, CSS properties or Html elements. Our empirical evaluation shows that X-Diag is effective in locating the root causes of cross-browser issues, and can provide useful support to developers for (eventually) eliminate XBIs.

关键词： Browsers Cascading style sheets Tools Engines Debugging Web pages Layout

来源：评论

学校读者我要写书评

暂无评论

STEED: An analytical database system for TrEE-structured data 43

STEED: An analytical database system for TrEE-structured dat...

引用

43rd International Conference on Very Large Data Bases, VLDB 2017

作者： Wang, Zhiyi Zhou, Dongyan Chen, Shimin State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences China

Tree-structured data formats, such as JSON and Protocol Buffers, are capable of expressing sophisticated data types, including nested, repeated, and missing values. While such expressing power contributes to their popularity in real-world applications, it presents a significant challenge for systems supporting tree-structured data. Existing systems have focused on general-purpose solutions either extending RDBMSs or designing native systems. However, the general-purpose approach often results in sophisticated data structures and algorithms, which may not reflect and optimize for the actual structure patterns in the real world. In this demonstration, we showcase Steed, an analytical database System for tree-structured data. We use the insights gained by analyzing representative real-world tree structured data as guidelines in the design of Steed. Steed learns and extracts a schema tree for a data set and uses the schema tree to reduce the storage space and improve the efficiency of data field accesses. We observe that sub-structures in real world data are often simple, while the treestructured data types can support very sophisticated structures. We optimize the storage structure, the column assembling algorithm, and the in-memory layout for the simple sub-structures (a.k.a. simple paths). Compared to representative state-of-the-art systems (i.e. PostgreSQL/JSON, MongoDB, and Hive+Parquet), Steed achieves orders of magnitude better performance for data analysis queries. © 2017 VLDB Endowment.

关键词： Database 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：