检索结果-内蒙古大学图书馆

Artificial Intelligence and Computer Engineering (ICAICE), International Conference on

作者： Yue Zhou Yue Yu Bo Ding Key Lab. of Parallel and Distributed Computing Lab. College of Computer National University of Defence Technology ChangSha China

ISBN: (数字)9781728191461

ISBN: (纸本)9781728191478

The development and deployment of machine learning (ML) applications differ significantly from traditional applications in many ways, which have led to an increasing need for efficient and reliable production of ML applications and supported infrastructures. Though platforms such as TensorFlow Extended (TFX), ModelOps, and Kubeflow have provided end-to-end lifecycle management for ML applications by orchestrating its phases into multistep ML pipelines, their performance is still uncertain. To address this, we built a functional ML platform with DevOps capability from existing continuous integration (CI) or continuous delivery (CD) tools and Kubeflow, constructed and ran ML pipelines to train models with different layers and hyperparameters while time and computing resources consumed were recorded. On this basis, we analyzed the time and resource consumption of each step in the ML pipeline, explored the consumption concerning the ML platform and computational models, and proposed potential performance bottlenecks such as GPU utilization. Our work provides a valuable reference for ML pipeline platform construction in practice.

关键词： Training Computational modeling Pipelines Graphics processing units Tools Data models Task analysis

来源：评论

学校读者我要写书评

暂无评论

Device View Redundancy: an adaptive low-overhead fault tolerance mechanism for many-core system

Device View Redundancy: an adaptive low-overhead fault toler...

引用

International Workshop on Intelligent Communication and Social Networks

作者： Wentao Jia Chunyuan Zhang Jian Fu National Key Laboratory of Parallel and Distributed Processing College of Computer National University of Defense Technology Institute for Informatics University of Amsterdam

ISBN: (纸本)9781479909735

Continued increasing of fault rate in integrate circuit makes processors more susceptible to errors, especially many-core processor. Meanwhile, most systems or applications do not need full fault coverage, which has excessive overhead. So on-demand fault tolerance is desired for these applications. In this paper, we propose an adaptive low-overhead fault tolerance mechanism for many-core system, called Device View Redundancy (DVR). It treats fault tolerance as a device that can be configured and used by application when high reliability is needed. Nevertheless, DVR exploits the idle resources for low-overhead fault tolerance, which is based on the observation that the utilization of many-core system is low due to lack of parallelism in application. Finally, the experiment shows that the performance overhead of DVR is reduced by 16% to 98% compared with full Dual Modular Redundancy (DMR).

关键词： On-demand redundancy Idle resource exploitation Dynamic coupling Low-overhead Many core system

来源：评论

学校读者我要写书评

暂无评论

IWFR: Exploiting full-duplex in wireless relay networks

IWFR: Exploiting full-duplex in wireless relay networks

引用

International Conference on Information Science and Technology (ICIST)

作者： Yong Lu Shaohe Lv Xiaodong Wang Xingming Zhou National Key Laboratory for Parallel and Distributed Processing School of Computer Science National University of Defense Technology Changsha Hunan China

Traditional wireless relay networks have large end-to-end time delay and low throughput because of the limit that it can't receive and forward at the same time. In this paper, we proposed IWFR: Immediate Wireless Full-Duplex Relay which exploits the advantages of full-duplex to shorten the end-to-end time delay and improve the throughput. At the same time, we designed a new implicit acknowledgement mechanism, which can eliminate the ACK overheads and evidently improve the throughput of the relay. To implement IWFR, we also modified the full-duplex node architecture to make it support for immediate relay. Simulation shows that IWFR shortens the end-to-end time delay by 60% on average and improves the throughput to 240% of the original relay.

关键词： Wireless communication Throughput Delay effects Wireless sensor networks Interference Relay networks (telecommunications)

来源：评论

学校读者我要写书评

暂无评论

An energy-efficient scheduling algorithm for sporadic real-time tasks in multiprocessor systems

An energy-efficient scheduling algorithm for sporadic real-t...

引用

13th IEEE International Workshop on FTDCS 2011, the 8th International Conference on ATC 2011, the 8th International Conference on UIC 2011 and the 13th IEEE International Conference on HPCC 2011

作者： Zhang, Dong-Song Chen, Fang-Yuan Li, Hong-Hua Jin, Shi-Yao Guo, De-Ke National Laboratory for Parallel and Distributed Processing School of Computer National University of Defense Technology Changsha 410073 China School of Computing Science Simon Fraser University 8888 University Drive Burnaby BC V5A 1S6 Canada National Laboratory for Information System Engineering School of Information Systems and Management National University of Defense Technology Changsha 410073 China

ISBN: (纸本)9780769545387

As the energy consumption of embedded multiprocessor systems becomes increasingly prominent, the real-time energy-efficient scheduling in multiprocessor systems becomes an urgent problem to reduce the system energy consumption while meeting real-time constraints. For a multiprocessor with independent DVFS and DPM at each processor, this paper proposes an energy-efficient real-time scheduling algorithm named LRE-DVFS-EACH, based on LRE-TL which is an optimal real-time scheduling algorithm for sporadic tasks. LRE-DVFS-EACH utilizes the concept of TL plane and the idea of fluid scheduling to dynamically scale the voltage and frequency of processors at the initial time of each TL plane as well as the release time of a sporadic task in each TL plane. Consequently, LRE-DVFS-EACH can obtain a reasonable tradeoff between the real-time constraints and the energy saving. LRE-DVFS-EACH is also adaptive to the change of workload caused by the dynamic release of sporadic tasks, which can obtain more energy savings. The experimental results show that compared with existing algorithms, LRE-DVFS-EACH can not only guarantee the optimal feasibility of sporadic tasks, but also achieve more energy savings in all cases, especially in the case of high workloads. © 2011 IEEE.

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

EmbRace: Accelerating Sparse Communication for distributed Training of NLP Neural Networks

arXiv

引用

arXiv 2021年

作者： Li, Shengwei Lai, Zhiquan Li, Dongsheng Zhang, Yiming Ye, Xiangyu Duan, Yabo National Key Laboratory of Parallel Distributed Processing Computer College National University of Defense Technology China Xiamen University China

distributed data-parallel training has been widely adopted for deep neural network (DNN) models. Although current deep learning (DL) frameworks scale well for dense models like image classification models, we find that these DL frameworks have relatively low scalability for sparse models like natural language processing (NLP) models that have highly sparse embedding tables. Most existing works overlook the sparsity of model parameters thus suffering from significant but unnecessary communication overhead. In this paper, we propose EmbRace, an efficient communication framework to accelerate communications of distributed training for sparse models. EmbRace introduces Sparsity-aware Hybrid Communication, which integrates AlltoAll and model parallelism into data-parallel training, so as to reduce the communication overhead of highly sparse parameters. To effectively overlap sparse communication with both backward and forward computation, EmbRace further designs a 2D Communication Scheduling approach which optimizes the model computation procedure, relaxes the dependency of embeddings, and schedules the sparse communications of each embedding row with a priority queue. We have implemented a prototype of EmbRace based on PyTorch and Horovod, and conducted comprehensive evaluations with four representative NLP models. Experimental results show that EmbRace achieves up to 2.41X speedup compared to the state-of-the-art distributed training baselines. Copyright © 2021, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Antenna calibration method for MMSE-based network MIMO system

Antenna calibration method for MMSE-based network MIMO syste...

引用

2015 IEEE/CIC International Conference on Communications in China - Workshops, CIC/ICCC 2015

作者： Liu, Qiang Su, Xin Zeng, Jie Liu, Lili Lv, Tiejun Tsinghua National Laboratory for Information Science and Technology Research Institute of Information Technology Tsinghua University Beijing China Key Laboratory of Trustworthy Distributed Computing and Service School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing China

ISBN: (纸本)9781467397018

In this paper, we propose a new channel reciprocity calibration method based on minimum mean square error (MMSE) in network multiple-input and multiple-output (MIMO) system. Network MIMO system can mitigate the inter-cell interference by joint precoding techniques. The system is assumed to operate in time division duplex (TDD). However, the generalized channeles do not only consist of free-space electromagnetic channel, but also include transceiver modules, the mismatches between the transceiver modules of the antennas destroy the channel reciprocity. Therefore the mismatches should be calibrated before channel reciprocity can be utilized. An effective method for calculating the calibration coefficients in the network MIMO system whose BSs are equipped with multiple antennas is proposed. The network MIMO system is divided into M independent centralized MIMO systems and one distributed MIMO system, then the calibration coefficients of the two systems can be combined to calculate the calibration coefficients of the network MIMO system. The validity of the method is proved by the simulation results. © 2015 IEEE.

关键词： Mean square error

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Broker Approach to Semantics-Based Service Composition

An Intelligent Broker Approach to Semantics-Based Service Co...

引用

IEEE Annual International Computer Software and Applications Conference (COMPSAC)

作者： Yufeng Zhang Hong Zhu National Laboratory for Parallel and Distributed Processing Department of Computing Science National University of Defense Technology Changsha China Department of Computing and Electronics Oxford Brookes University Oxford UK

This paper proposes an intelligent broker approach to service composition and collaboration. The broker employs a planner to generate service composition plans according to service usage and workflow knowledge, dynamically searches for services according to the plan, then invokes and coordinates the executions of the selected services at runtime. A prototype called I-Broker has been implemented to support the approach, which can be instantiated by populating the knowledge-base with domain specific knowledge to form domain specific brokers. This paper also reports experiments that evaluate the scalability of the approach.

关键词： Knowledge based systems Ontologies Semantics Complexity theory Planning Scalability Prototypes

来源：评论

学校读者我要写书评

暂无评论

Property-oriented testing of real-time systems

Property-oriented testing of real-time systems

引用

Asia-Pacific Conference on Software Engineering

作者： Shuhao Li Ji Wang Wei Dong Zhi-Chang Qi National Laboratory for Parallel and Distributed Processing Changsha China State Key Laboratory for Software Engineering Wuhan University of China China

Although statecharts has gained widespread use as a formalism for modeling reactive real-time systems, testing these systems still confronts some difficulties, of which a major one is the existence of numerous and complex system behaviors. It is extremely difficult to conduct comprehensive and in-depth testing of such real-time systems. This paper presents an approach to property-oriented real-time testing. Necessary real-time extensions are proposed such that the time-enriched statecharts can describe nontrivial timing constraints. The properties to be tested are characterized by a restricted real-time logic. Then the targeted test sequences are derived from the real-time models according to the user-specified properties. Using this approach, testing efforts can be focused on particular properties of the real-time systems and usually only a small portion of the total behaviors needs to be tested.

关键词： System testing Real time systems Software testing Software engineering

来源：评论

学校读者我要写书评

暂无评论

A Connectivity-Enhanced Multi-Task Learning based on Anatomical Priors for 3D Class-Balanced Pulmonary Airway Segmentation

A Connectivity-Enhanced Multi-Task Learning based on Anatomi...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Yan Jia Yong Dou Peng Qiao Yuqing Cheng Kele Xu Zhouyu He National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha China College of Systems Engineering National University of Defense Technology Changsha China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Accurate and efficient airway segmentation is essential for evaluating pulmonary diseases, aiding diagnosis, reducing the preoperative burden of airway identification, and minimizing patient discomfort during prolonged surgeries. However, current pulmonary airway reconstruction techniques are hindered by two major challenges: difficulty in accurately reconstructing fine airway branches due to the tendency to overlook small targets, and insufficient structural connectivity leading to frequent branch discontinuities within the airway tree. These limitations directly affect the clinical applicability of reconstructed airways. To overcome these challenges, a novel 3D pulmonary airway segmentation multi-task framework is proposed, designed to enhance the performance of existing backbone models. This approach integrates Anatomical Prior-Based Multi-Task Learning (AP-MTL) through the use of Gaussian-constructed connectivity-enhanced isosurfaces, significantly improving the network’s ability to maintain airway continuity. Additionally, a Class-Balanced CT Density Distribution Reconstruction mechanism (DDR-CB) is introduced, further refining the model’s capability to detect and segment fine airway branches. As a result of these enhancements, the model demonstrates a 11.5% average improvement in segmentation accuracy and connectivity compared to the baseline. The source code is publicly accessible at https://***/inexhaustible419/APMTLAirwaySegment.

关键词： Image segmentation Three-dimensional displays Accuracy Lungs Atmospheric modeling Supervised learning Surgery Multitasking Image reconstruction Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

Named Entity Recognition of Chinese Diabetic Literature by Integrating General Domain Knowledge 4

Named Entity Recognition of Chinese Diabetic Literature by I...

引用

4th IEEE International Conference on Computer and Communication Engineering Technology, CCET 2021

作者： Zhang, Duochuan Beijing University of Posts and Telecommunications School of Computer National Demonstrative Software School Beijing China Key Laboratory of Trustworthy Distributed Computing and Service Beijing University of Posts and Telecommunications Ministry of Education Beijing China

ISBN: (纸本)9781665438902

In the field of named entity recognition of diabetes literature, due to the lack of large-scale and high-quality annotation data sets, the traditional named entity recognition technology has the drawback of insufficient language feature extraction and poor generalization ability. In view of the above problems, this paper proposes an entity recognition method that integrates general domain knowledge for naming entity recognition in diabetes literature. This method uses Word2vec algorithm to generate feature level vectors based on Chinese diabetes literature. On this basis, the feature vectors are fused with the feature vectors generated by the pretraining model of BERT on the Chinese Wikipedia corpus, and uses principal component analysis to reduce the redundant feature vectors and enhance the expression ability of feature vectors in the general language scope. Then the context features are extracted by bidirectional long-term and short-term memory network, and finally the entity label sequence is generated by conditional random field decoding. The experiment proves that the model which integrates general domain knowledge is better than other traditional models in the recognition of Chinese diabetes literature, and the highest F1 value is 71.08%. © 2021 IEEE.

关键词： Domain Knowledge

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：