检索结果-内蒙古大学图书馆

Development of high power SiC devices for rail traction power systems

JOURNAL OF CRYSTAL GROWTH 2019年 507卷 442-452页

作者： Liu, Guoyou Wu, Yibo Li, Kongjing Wang, Yangang Li, ChengZhan State Key Lab Adv Power Semicond Devices Zhuzhou Hunan Peoples R China Dynex Semicond Ltd Power Semicond R&D Ctr Lincoln England

Si-based power devices have been widely used in electric locomotive and EMU, however, there are urgent requirements from the industry to have power converters with smaller size and higher performance. To meet such demands, wide band-gap (WBG) devices such as SiC power chips and modules are developed for the traction systems, and a 1.7 kV hybrid SiC power module has been used in Metro system and all SiC 3.3 kV power module has been developed. In this paper, the development of hybrid SiC and all SiC power modules are introduced from the perspectives of devices, modules and applications. SiC SBD and MOSFET chips and modules are manufactured and tested, both static and dynamic testing results are evaluated and showed that SiC devices present better performance and higher efficiency than the traditional Si IGBT modules, especially in the conditions of high temperature and high operation frequency. Based on the developed SiC modules, the urgent demand of traction converters with smaller size and lighter weight can be achieved while present advantages such as higher efficiency, power density, operation temperatures and frequency. The current challenges of SiC devices and future applications are discussed for the next generation of rail transportation.

关键词： SiC device Rail traction computer simulation Heat transfer Semiconducting silicon

来源：评论

学校读者我要写书评

暂无评论

A Learning-Exploring Method to Generate Diverse Paraphrases with Multi-Objective Deep Reinforcement Learning 28

A Learning-Exploring Method to Generate Diverse Paraphrases ...

引用

28th International conference on Computational Linguistics, COLING 2020

作者： Liu, Mingtong Yang, Erguang Xiong, Deyi Zhang, Yujie Meng, Yao Hu, Changjian Xu, Jinan Chen, Yufeng School of Computer Science and Information Technology Beijing Key Lab of Traffic Data Analysis and Mining Beijing Jiaotong University Beijing China College of Intelligence and Computing Tianjin University Tianjin China Lenovo Research AI Lab. Beijing China

ISBN: (纸本)9781952148279

Paraphrase generation (PG) is of great importance to many downstream tasks in natural language processing. Diversity is an essential nature to PG for enhancing generalization capability and robustness of downstream applications. Recently, neural sequence-to-sequence (Seq2Seq) models have shown promising results in PG. However, traditional model training for PG focuses on optimizing model prediction against single reference and employs cross-entropy loss, which objective is unable to encourage model to generate diverse paraphrases. In this work, we present a novel approach with multi-objective learning to PG. We propose a learning-exploring method to generate sentences as learning objectives from the learned data distribution, and employ reinforcement learning to combine these new learning objectives for model training. We first design a sample-based algorithm to explore diverse sentences. Then we introduce several reward functions to evaluate the sampled sentences as learning signals in terms of expressive diversity and semantic fidelity, aiming to generate diverse and high-quality paraphrases. To effectively optimize model performance satisfying different evaluating aspects, we use a GradNorm-based algorithm that automatically balances these training objectives. Experiments and analyses on Quora and Twitter datasets demonstrate that our proposed method not only gains a significant increase in diversity but also improves generation quality over several state-of-the-art baselines. © 2020 COLING 2020 - 28th International conference on Computational Linguistics, Proceedings of the conference. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

iMODE (interactive MOod Detection Engine) Processor 4

iMODE (interactive MOod Detection Engine) Processor

引用

4th International conference on computer Science and Engineering (UBMK)

作者： Savas, Whammed Emin Guney, Isa Ahmet Tokatli, Nazli Nakeeb Kisinbay, Berk Kucuk, Gurhan Yeditepe Univ Dept Comp Engn Istanbul Turkey

ISBN: (纸本)9781728139647

Applications change their mood from time to time. A memory-intensive application does not need to be always memory-intensive from its first instruction to its last. Similarly, it is highly usual that a computation-intensive application generates heavy memory traffic at certain points of its entire run. Meanwhile, processors are designed to have a fixed resource configuration, which is expected to serve all kinds of applications with all kinds of program phases. In this paper, we propose the iMODE processor, which tracks down instant mood changes of running applications and apply immediate processor mode changes between in-order and out-of-order modes for either saving power or keeping up with the high-performance demands of applications. In our experiments, we show that the iMODE processor can really track the mood changes of applications in an accurate manner, and achieves 17% power savings with only less than 1% of performance drop, on the average across all simulated benchmarks.

关键词： computer architecture reconfigurable architectures energy consumption

来源：评论

学校读者我要写书评

暂无评论

A new storage workload management model for high-performance I/O systems 21

A new storage workload management model for high-performance...

引用

21st IEEE International conference on high performance Computing and Communications, 17th IEEE International conference on Smart City and 5th IEEE International conference on Data Science and Systems, HPCC/SmartCity/DSS 2019

作者： Huang, Huang Sheng, You Yutong, Lu College of Computer National University of Defense Technology Changsha 410073 China National Supercomputer Center in Changsha Changsha410082 China School of Data and Computer Science Sun Yat-Sen University Guangzhou510006 China

ISBN: (纸本)9781728120584

Data-intensive applications involve with massive data, ranging from terabytes to petabytes, which is a critical bottleneck in the next generation systems. In this article, we mainly focus on system I/O competition caused by the data-intensive applications and unbalanced storage capacity, and attempt to develop effective solutions to alleviate this issue. To this end, we develop an I/O middleware based method. The core of our middleware is the storage workload management model (SWMM), which allows us to adjust and optimize the I/O access process efficiently and flexibly. We have implemented our solution and conducted extensive tests based on a petascale system. The experimental results demonstrate the effectiveness and efficiency of our solution. © 2019 IEEE.

关键词： Middleware

来源：评论

学校读者我要写书评

暂无评论

A Scalable Architecture for CNN Accelerators Leveraging high-performance Memories

A Scalable Architecture for CNN Accelerators Leveraging High...

引用

IEEE conference on high performance Extreme Computing (HPEC)

作者： Maarten Hattink Giuseppe Di Guglielmo Luca P. Carloni Keren Bergman Columbia University in the City of New York New York USA

ISBN: (数字)9781728192192

ISBN: (纸本)9781728192208

As FPGA-based accelerators become ubiquitous and more powerful, the demand for integration with high-performance Memory (HPM) grows. Although HPMs offer a much greater bandwidth than standard DDR4 DRAM, they introduce new design challenges such as increased latency and higher bandwidth mismatch between memory and FPGA cores. This paper presents a scalable architecture for convolutional neural network accelerators conceived specifically to address these challenges and make full use of the memory's high bandwidth. The accelerator, which was designed using high-level synthesis, is highly configurable. The intrinsic parallelism of its architecture allows near-perfect scaling up to saturating the available memory bandwidth.

关键词： Random access memory Bandwidth computer architecture Parallel processing Convolutional neural networks Standards Field programmable gate arrays

来源：评论

学校读者我要写书评

暂无评论

Female computer Scientists Needed: Approaches For Closing The Gender Gap

Female Computer Scientists Needed: Approaches For Closing Th...

引用

Frontiers in Education (FIE) conference

作者： Corinna Kröhn Iris Groher Barbara Sabitzer Lisa Kuka Johannes Kepler University Linz Austria

ISBN: (数字)9781728189611

ISBN: (纸本)9781728189628

Despite great effort in research and teaching as well as in raising women's quotas in the public sector and in enterprises, a gender gap in computer science can still be observed. The reasons are manifold, including the lack of interest of girls because of existing stereotypes of male nerds, misconceptions in the field of computer science, and concrete differences in students' self-concept and performance. Girls do not choose schools or studies in the field of computer science because working on computers or coding is unattractive to them. Research shows poorer performance, missing interest, and lower self-concept for girls in secondary computing education as well as a high dropout rate of female students in computer science courses at the bachelor's level. Software development exams contribute to those high dropout rates, and women often choose a different subject or even stop their university education altogether. This leads to fewer female employees in the field of computer science, and this situation calls for special measures at several levels. The current paper provides an overview of the initiatives of our university that aim to address several aspects of the gender gap. The programs offered aim to (1) increase interest in computer science and recruit (highly) gifted girls for our talent programs, (2) recruit and support female bachelor's and master's students in computer science, and (3) redesign programming courses and teaching materials to reduce the gender gap in performance and support, especially for female students. This study also describes a "cyber tutoring" program in which highly gifted girls between 13 and 16 years collaborate with and are supervised by female role models in higher positions in STEM. The evaluation results gained in the first year of this case study are promising and suggest that the extension and further development of the program would be advantageous.

关键词： computer science Manifolds Current measurement Education Software Encoding programming profession

来源：评论

学校读者我要写书评

暂无评论

I/O scheduling for limited-size burst-buffers deployed high performance computing 20

I/O scheduling for limited-size burst-buffers deployed high ...

引用

20th International conference on Parallel and Distributed Computing, Applications and Technologies, PDCAT 2019

作者： Zha, Benbo Shen, Hong School of Data and Computer Science Sun Yat-sen University Guangzhou China School of Computer Science University of Adelaide Adelaide Australia

ISBN: (纸本)9781728126166

Burst-Buffers is a high throughput, small size intermediate storage system integrated between computing nodes and permanent storage system to mitigate the I/O bottleneck problem in modern high performance Computing (HPC) platforms. This system, however, is unable to effectively handle variable-intensity I/O bursts resulted by unpredictable concurrent accesses to the shared Parallel File System (PFS). In this paper, we introduce a probabilistic I/O scheduling method that takes into account of the burst-buffer load state and instantaneous I/O load distribution of the system based on the probabilistic model of applications to relieve the I/O congestion when I/O load exceeds the PFS bandwidth caused by dynamic application interference. The proposed scheduling method for limited-size Burst-Buffers deployed HPC platforms makes online decision of probabilistic selection of concurrent I/O requests for going through (to PFS), buffering (to Burst-Buffers) or declination in accordance to both the available I/O bandwidth and the current buffer state in order to maximize system efficiency or minimize application dilation. Extensive experiment results on actual characteristic synthetic data show that our method handles the I/O congestion effectively. © 2019 IEEE.

关键词： Scheduling

来源：评论

学校读者我要写书评

暂无评论

Band-Pass Filtering for Non-Stationary Noise in Agricultural Images to Pest Control Based On Adaptive Semantic Modeling

Band-Pass Filtering for Non-Stationary Noise in Agricultural...

引用

International conference on Semantic Computing

作者： Alex B. Bertolla Paulo E. Cruvinel Post-Graduation Program in Computer Science Federal University of São Carlos SP Brazil

ISBN: (数字)9781728188997

ISBN: (纸本)9781728189000

Image analysis has been used in a very large scale for different purposes. When an image is captured by a digital sensor, it is usually affected by some type of noise, even the smoothest ones. Therefore, image enhancement and denoising process are important tasks of digital image processing. This paper presents an algorithm to reduce non-stationary noise with the combination of a Low-Pass Filter (LPF) and a high-Pass Filter (HPF), in conjunction with an adaptive semantic model. To simulate the usefulness of such arrangement, a non-stationary Gaussian noise has been applied to an image, which has been splitted into the four quadrants, all of them having the same dimensions. In fact, such a noise with different intensities, has been added to the image in each of its quadrants. The Peak Signal-to-Noise Ratio (PSNR) has been used to measure the best cutoff frequencies for both filters, as well as rules based on semantic concepts have been structured for decision making. Furthermore, for the validation of the algorithm we have taken into account the evaluation of the Mean Squared Error (MSE) using a typical digital image obtained from a crop of maize with the presence of the earwornm (Helicoverpa Zea). Besides, the denoising process demonstrates the efficiency and the satisfactory performance for the non-stationary noise filtering in agricultural images.

关键词： Band-pass filters PSNR Semantics Cutoff frequency Low-pass filters Filtering algorithms Digital filters

来源：评论

学校读者我要写书评

暂无评论

high performance radar images modelling and recognition of real objects 5

High performance radar images modelling and recognition of r...

引用

Data Science Session at the 5th International conference on Information Technology and Nanotechnology, DS-ITNT 2019

作者： Zherdev, D.A. Prokudin, V.V. Samara National Research University Moskovskoe Shosse 34А Samara443086 Russia Image Processing Systems Institute of RAS Branch of the FSRC "Crystallography and Photonics" RAS Molodogvardejskaya street 151 Samara443001 Russia

In the work there is a modernization of the parallel algorithm for the radar images formation of 3D models with the synthesis of the antenna aperture. In the formation of the scene description, the various structures are used in which it is possible to use more efficient and derived calculations. In addition, it is the topical task to recognize objects on radar images. Thus, on the basis of the implemented parallel program for modelling, the high performance required for simulating multiple radar images can be achieved. © 2019 CEUR Workshop Proceedings.

关键词： Radar imaging

来源：评论

学校读者我要写书评

暂无评论

Trust-based distributed kalman filter estimation fusion under malicious cyber attacks 21

Trust-based distributed kalman filter estimation fusion unde...

引用

作者： Chen, Yanping Ma, Long Xia, Hong Gao, Cong Wang, Zhongmin Yu, Zhong School of Computer Science AND Technology Xi'an University of Posts and Telecommunications Xi'an710121 China Shananxi Key Laboratory of Network Data Analysis and Intelligent Processing Xi'an University of Posts and Telecommunications Xi'an710121 China School of Communications and Information Engineering Xi'an University of Posts and Telecommunications Xi'an710121 China

ISBN: (纸本)9781728120584

We consider distributed Kalman filter for dynamic state estimation over wireless sensor networks. It is promising but challenging when network is under cyber attacks. Since the information exchange between nodes, the malicious attacks quickly spread across the entire network, which causing large measurement errors and even to the collapse of sensor networks. Aiming at the malicious network attack, a trust-based distributed processing frame is proposed. Which allows neighbor nodes to exchange information, and a series of trusted nodes are found using truth discovery. As a demonstration, distributed Cooperative Localization is considered, and numerical results are provided to evaluate the performance of the proposed approach by considering random, false data injection and replay attacks. © 2019 IEEE.

关键词： Kalman filters

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：