检索结果-内蒙古大学图书馆

11th EAI International Conference on Wireless and Satellite Systems, WiSATS 2020

作者： Zhang, Xiaojun Li, Na Li, Jun Chen, Chengguan Li, Hengzhong Chen, Geng Shandong University of Science and Technology Qingdao China State Key Laboratory of High-End Server and Storage Technology Jinan China

ISBN: (纸本)9783030690717

Compared with the SC decoder, BP decoder provides a higher throughput and lower decoding latency for its inherent parallel nature. However, the functional units of existing BP decoders are not fully utilized. In this paper, we propose a new update scheduling scheme and hardware optimization design to improve the hardware efficiency of BP decoder. First, a pipelined decoder architecture is proposed to reduce the consumption of functional units. Then, a new update scheduling scheme is proposed, when updating the messages, both new-value and old-value approaches are used to improve the utilization of functional units and reduce the decoding latency. The analysis and synthesis results have shown that, compared with the existing methods, the proposed decoder suffers from a slight the decoding performance degradation, but the utilization rate of basic computational blocks (BCB) can be increased to 50.58%. The storage resource dissipation can be reduced by 20.1%–34.09%. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Belief propagation

来源：评论

学校读者我要写书评

暂无评论

A Low-complexity Max Unpooling Achitecture for CNNs

A Low-complexity Max Unpooling Achitecture for CNNs

引用

International Conference on ASIC

作者： Xiaojun Zhang Chenshi Zhu Qin Han Zhengrong Wang Dexue Zhang College of Electronic and Information Engineering Shandong University of Science and Technology State Key Laboratory of High-end Server and Storage Technology Qingdao China

In the training of Convolutional Neural Networks (CNN), the conventional max unpooling architecture introduces significant sparsity for zero-filling during upsampling, leading to additional resource costs in adjacent deconvolution layers. In order to leverage the high sparsity of the recovery matrix, this paper proposes a low-complexity hardware architecture. This architecture solely stores the input matrix in the max unpooling layer, eliminating the need for the upsampling recovery process, while mapping the stored relative indices from the forward inference to deconvolution. Moreover, the deconvolution process exclusively involves non-zero elements, reducing computational and storage requirements. This paper derives the distribution of the maximum number of non-zero values corresponding to the deconvolution window for different layers, demonstrating its compatibility with networks of various sizes. Synthesizing the proposed architecture on Arria 10 AX115N2F45E1SG, for the VGG16, it can save 60.7% of ALMs and 68.8% of storage resources, Moreover, the proposed architecture enhances the frequency by 8.0%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Genetic Algorithm based hyper-parameters optimization for transfer Convolutional Neural Network

Genetic Algorithm based hyper-parameters optimization for tr...

引用

2022 International Conference on Advanced Algorithms and Neural Networks, AANN 2022

作者： Li, Chen Jiang, JinZhe Zhao, YaQian Li, RenGang Wang, endong Zhang, Xin Zhao, Kun State Key Laboratory of High-end Server & Storage Technology Beijing100085 China Inspur Group Co. Ltd Beijing100085 China Shandong Hailiang Information Technology Institutes Shandong Jinan250014 China State Key Laboratory of High-End Server & Storage Technology Shandong Jinan250014 China Inspur Electronic Information Industry Co. Ltd Shandong Jinan250014 China Guangdong Inspur Intelligent Computing Technology Co. Ltd Guangdong Guangzhou510000 China

ISBN: (纸本)9781510655751

Hyperparameter optimization is a challenging problem in developing deep neural networks. Decision of transfer layers and trainable layers is a major task for design of the transfer convolutional neural networks (CNN). Conventional transfer CNN models are usually manually designed based on intuition. In this paper, a genetic algorithm is applied to select trainable layers of the transfer model. The filter criterion is constructed by accuracy and the counts of the trainable layers. The results show that the method is competent in this task. The system will converge with a precision of 97% in the classification of Cats and Dogs datasets, in no more than 15 generations. Moreover, backward inference according the results of the genetic algorithm shows that our method can capture the gradient features in network layers, which plays a part on understanding of the transfer AI models. © 2022 SPIE.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Research and Implementation of DMA Based on SR-IOV

Research and Implementation of DMA Based on SR-IOV

引用

Mobile Internet, Cloud Computing and Information Security (MICCIS), International Conference on

作者： Hongliang Wang Peiqiang Dong Qi Mu Wei Liu Shengcai Lu Runze Zhang State Key Laboratory of High-End Server and Storage Technology Inspur Electronic Information Industry Co. Ltd Jinan China

With the development of cloud computing, container cloud, and other technologies, CPU and memory virtualization becomes more mature. Since the performance of I/O virtualization affects the overall operational efficiency of the system, improving I/O performance and the utilization of transmission bandwidth of devices is the focus of virtualization technology research. Based on the theoretical research of Single Root I/O Virtualization (SR-IOV) technology, this paper designs and implements a Direct Memory Access(DMA) interaction architecture based on SR-IOV virtualization devices using Field Programmable Gate Array (FPGA), which solves the need for high-speed data transmission between cloud computing platforms and virtual devices. This architecture adopts the processing method of independent virtual device management descriptors and shared data movers, ensuring independent operation between virtual devices and reducing the resource utilization rate of the design. Firstly, the chain structure is employed flexibly and efficiently to achieve descriptor management. In the meantime, it also reduces the pressure of interrupts on the CPU. Secondly, the data mover implementation framework adopts asynchronous pipelining mode, with extremely high transmission performance. After testing and verification, the interactive architecture can achieve high-speed data communication, with read and write bandwidth utilization rates of up to 87% and 79%, respectively. At the same time, this architecture can balance the bandwidth allocation among virtual devices.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design of Fast-SSC Decoder for STT-MRAM Channel 11th

Design of Fast-SSC Decoder for STT-MRAM Channel

引用

11th EAI International Conference on Wireless and Satellite Systems, WiSATS 2020

作者： Cui, Jianming Bao, Zengxiang Zhang, Xiaojun Guo, Hua Chen, Geng Shandong University of Science and Technology Qingdao266590 China State Key Laboratory of High-End Server and Storage Technology Jinan China

ISBN: (纸本)9783030690717

In order to achieve fast decoding and improve the throughput, this paper uses the polar code to encode the spin transfer torque MARAM (STT-MRAM) channel. Based on the Fast-SSC algorithm, a (256, 220) hardware architecture is designed, including the controller, processing element, Kronecker product and memory module. This paper reduces the complexity of data process by splitting the data of nodes, and reduces the memory bandwidth by increasing the reusability of data. The decoder is synthesized on Stratix V 5SGXEA7N2F45C2, the decoding latency is 0.68us, and it can achieve 375 Mbps at 167 MHz. © 2021, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.

关键词： Reusability

来源：评论

学校读者我要写书评

暂无评论

Modeling and Characteristic Analysis of Clutter for Airborne Emitter Non-cooperative Detection System 12th

Modeling and Characteristic Analysis of Clutter for Airborne...

引用

12th International Conference on Computer Engineering and Networks, CENet 2022

作者： Song, Jie Lou, Shu-li Gong, Wei-feng Research Institute of Information Fusion Naval Aviation University Yantai China School of Physics and Electronic Information Yantai University Yantai China State Key Laboratory of High-End Server and Storage Technology Jinan China

ISBN: (纸本)9789811969003

In the configuration scenario of bistatic radar, clutter tends to be complicated, which will be aggravated by the moving state of non-cooperative radiators. Therefore, a clutter modeling method of non-cooperative airborne radiators is proposed. Based on the background of non-cooperative radar emitter and ground receiving station, the clutter spectrum under the system is constructed and its distribution is studied theoretically. Simulation results show that the influence of clutter on radar decreases with the increase of distance, when the radiation source location and velocity are different, the greater the clutter freedom is, the greater the complexity of clutter spectrum is. The clutter modeling method of non-cooperative moving emitter extends the clutter model and lays a foundation for clutter suppression and target detection of non-cooperative emitter radar configuration. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Radar clutter

来源：评论

学校读者我要写书评

暂无评论

Sequence-to-Sequence Knowledge Graph Completion Based on Gated Attention Unit 29

Sequence-to-Sequence Knowledge Graph Completion Based on Gat...

引用

29th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2023

作者： Yi, Fengge Wei, Xiumei Liu, Xiaojing Jiang, Xuesong Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China State Key Laboratory of High-end Server & Storage Technology Jinan China

ISBN: (纸本)9798350330717

We present GauKGT5, a sequence-to-sequence model proposed for knowledge graph completion (KGC). Our research extends the KGT5 model, a recent sequence-to-sequence link prediction (LP) model. GauKGT5 takes advantage of textual characteristics inherent in the knowledge graph, exhibiting a small model size. However, KGT5's proficiency in link prediction necessitates the ensemble with a knowledge graph embedding model, which itself poses challenges due to its substantial size and expense. By integrating the Gated Attention Unit into the KGT5 model and directly applying it to the encoder-decoder structure, we achieve improved contextual dependency capturing within the sequence, resulting in enhanced prediction accuracy, accelerated training speed, and enhanced computational efficiency. At the same time, we introduce parallel computing as a means to enhance the efficiency of model training and inference within the XPU distributed computing environment. © 2023 IEEE.

关键词： gated attention unit knowledge graph completion sequence-to-sequence model

来源：评论

学校读者我要写书评

暂无评论

Bits Mapping in Triple-level-cell (TLC) Charge-trap (CT) 3D NAND Flash Memory and its Applications to IoT Security

Bits Mapping in Triple-level-cell (TLC) Charge-trap (CT) 3D ...

引用

2021 IEEE International Conference on Integrated Circuits, Technologies and Applications, ICTA 2021

作者： Xi, Yifang Fang, Xiaotong Kong, Yachen Guo, Yifan Lin, Hongzhe Zhan, Xuepeng Chen, Jiezhi Shandong University School of Information Science and Engineering China State Key Laboratory of High-end Server and Storage Technology China

ISBN: (纸本)9781665417471

Error bits distributions in Triple-level-cell (TLC) charge-trap (CT) 3D NAND flash memory chips are studied systematically in this work. It is found that, firstly, hard error bits exist without dependences on the data patterns;secondly, soft error bits can be observed when programming a certain data pattern, and the distributions are unique in each block;thirdly, random errors can be distinguished by using shift-read operations, which originates from the random telegraph noise in NAND string current. More importantly, by errors mapping technologies, memory chips' fingerprints could be generated, which can be utilized to assure data security. © 2021 IEEE.

关键词： Mapping

来源：评论

学校读者我要写书评

暂无评论

Bilateral Anomaly Detection Based on Edge Computing: Baseboard Management Controller

Bilateral Anomaly Detection Based on Edge Computing: Baseboa...

引用

International Symposium on Computational Intelligence and Design, ISCID

作者： Deqiang Zou Hongtao Man Tuo Li State Key Laboratory of High-End Server & Storage Technology Beijing China Inspur Electronic Information Industry Co. Beijing China

Anomaly detection on servers is a crucial technique with the increment of computing patterns in this period. To accommodate current demands, we propose a version of a bilateral detection scheme based on the baseboard management controller with machine learning to conduct compound anomaly detection and boost the upper bound. On the local part, the naive logic trees are deployed on the calculating module to make initial detection and carry out prompt actions maintaining the applicable working environment. Once the local anomaly alerts, the remote part is triggered to judge with the Gaussian Mixture Model, which ensures the anomaly type and proceeds with advanced steps. The bilateral detection system resorts to edge computing priority and scientifically allocates computing power to improve the level of anomaly detection on servers. The anomaly detection system’s effect herein is excellent in real-time demand and stability.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Swin transformer encoder-based StyleGAN for unbalanced endoscopic image enhancement

引用

Computers in Biology and Medicine 2024年 175卷 108472-108472页

作者： Deng, Bo Zheng, Xiangwei Chen, Xuanchi Zhang, Mingzhe School of Information Science and Engineering Shandong Normal University Jinan250352 China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan250352 China State Key Laboratory of High–end Server & Storage Technology Jinan250101 China

With the rapid development of artificial intelligence, automated endoscopy-assisted diagnostic systems have become an effective tool for reducing the diagnostic costs and shortening the treatment cycle of patients. Typically, the performance of these systems depends on deep learning models which are pre-trained with large-scale labeled data, for example, early gastric cancer based on endoscopic images. However, the expensive annotation and the subjectivity of the annotators lead to an insufficient and class-imbalanced endoscopic image dataset, and these datasets are detrimental to the training of deep learning models. Therefore, we proposed a Swin Transformer encoder-based StyleGAN (STE-StyleGAN) for unbalanced endoscopic image enhancement, which is composed of an adversarial learning encoder and generator. Firstly, a pre-trained Swin Transformer is introduced into the encoder to extract multi-scale features layer by layer from endoscopic images. The features are subsequently fed into a mapping block for aggregation and recombination. Secondly, a self-attention mechanism is applied to the generator, which adds detailed information of the image layer by layer through recoded features, enabling the generator to autonomously learn the coupling between different image regions. Finally, we conducted extensive experiments on a private intestinal metaplasia grading dataset from a Grade-A tertiary hospital. The experimental results show that the images generated by STE-StyleGAN are closer to the initial image distribution, achieving a Fréchet Inception Distance (FID) value of 100.4. Then, these generated images are used to enhance the initial dataset to improve the robustness of the classification model, and achieved a top accuracy of 86 %. © 2024 Elsevier Ltd

关键词： Image enhancement

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：