检索结果-南通市图书馆

Chinese Physics B 2014年第6期23卷 123-130页

作者：谢永华徐赵飞 Hans Burkhardt school of computer and software Nanjing University of Information Science and Technology Department of computer Science Freiburg University

This paper presents one novel spatial geometric constraints histogram descriptors （SGCHD） based on curvature mesh graph for automatic three-dimensional （3D） pollen particles recognition. In order to reduce high dimensionality and noise disturbance arising from the abnormal record approach under microscopy, the separated surface curvature voxels are ex- tracted as primitive features to represent the original 3D pollen particles, which can also greatly reduce the computation time for later feature extraction process. Due to the good invariance to pollen rotation and scaling transformation, the spatial geometric constraints vectors are calculated to describe the spatial position correlations of the curvature voxels on the 3D curvature mesh graph. For exact similarity evaluation purpose, the bidirectional histogram algorithm is applied to the spatial geometric constraints vectors to obtain the statistical histogram descriptors with fixed dimensionality, which is invariant to the number and the starting position of the curvature voxels. Our experimental results compared with the traditional methods validate the argument that the presented descriptors are invariant to different pollen particles geometric transformations （such as posing change and spatial rotation）, and high recognition precision and speed can be obtained simultaneously.

关键词： pollen recognition curvature mesh graph spatial geometric constraints bidirectional histogram

维普期刊数据库

An Application-Oriented Cache Allocation and Prefetching Method for Long-Running Applications in Distributed Storage Systems

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2019年第4期28卷 773-780页

作者： GUO Chang LI Ying LIU Hongzhi WU Zhonghai school of Electronics Engineering and computer Science Peking University National Engineering Research Center For software Engineering Peking University school of software and Microelectronics Peking University

Characteristics of long-running applications in cloud and big data environment are various and significantly influence the performance of cache *** gap between existing cache systems and the increasing performance requirements motivates us to propose the Application-oriented cache allocation and prefetching method(ACAP) to improve data access performance. An application-oriented cache allocation approach is designed based on hit count growth rates for a higher overall hit rate. Two application-oriented sequential prefetching approaches are proposed to improve the hit rate and prefetching accuracy by learning average read sizes of long-running applications. Based on correlation of data accesses, a parallelized correlated-directed prefetching approach is proposed to further increase the hit *** approaches are intergrated to obtain the maximized hit rate and prefetching accuracy. Experimental results on12 public real system traces show that ACAP achieves14.03%(up to 33.82%) higher prefetching accuracy and2.01%(up to 7.54%) higher hit rate compared with the best combination of baselines.

关键词： Cache prefetching Cache allocation Distributed storage system Resource management

Deep visual-linguistic fusion network considering cross-modal inconsistency for rumor detection

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2023年第12期66卷 16-32页

作者： Yang YANG Ran BAO Weili GUO De-Chuan ZHAN Yilong YIN Jian YANG school of computer Science and Engineering Nanjing University of Science and Technology National Key Laboratory for Novel software Technology Nanjing University school of software Shandong University

With the development of the Internet, users can freely publish posts on various social media platforms, which offers great convenience for keeping abreast of the world. However, posts usually carry many rumors, which require plenty of manpower for monitoring. Owing to the success of modern machine learning techniques, especially deep learning models, we tried to detect rumors as a classification problem automatically. Early attempts have always focused on building classifiers relying on image or text information, i.e., single modality in posts. Thereafter, several multimodal detection approaches employ an early or late fusion operator for aggregating multiple source information. Nevertheless, they only take advantage of multimodal embeddings for fusion and ignore another important detection factor, i.e., the intermodal inconsistency between modalities. To solve this problem, we develop a novel deep visual-linguistic fusion network(DVLFN) considering cross-modal inconsistency, which detects rumors by comprehensively considering modal aggregation and contrast information. Specifically, the DVLFN first utilizes visual and textual deep encoders, i.e., Faster R-CNN and bidirectional encoder representations from transformers, to extract global and regional embeddings for image and text modalities. Then, it predicts posts’ authenticity from two aspects:(1) intermodal inconsistency, which employs the Wasserstein distance to efficiently measure the similarity between regional embeddings of different modalities, and(2) modal aggregation, which experimentally employs the early fusion to aggregate two modal embeddings for prediction. Consequently, the DVLFN can compose the final prediction based on the modal fusion and inconsistency measure. Experiments are conducted on three real-world multimedia rumor detection datasets collected from Reddit, Good News, and Weibo. The results validate the superior performance of the proposed DVLFN.

关键词： multimodal learning Wasserstein distance rumor detection

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CYBERSPACE SECURITY

China Communications 2016年第11期13卷 68-69页

作者： Huanguo Zhang Yi Mu computer school in wuhan university school of computer science and software engineering centre for computer and information security research at university of wollongong

With rapid development and wide application of information technology,mankind is entering into an information *** world where people live and work has been changed to a brand new triple-dimensional space that consists of the physical world,the human society and the *** is well known that the cyberspace is an essential environment of human existence and development and it

关键词：网络空间安全人类社会信息技术信息时代三维空间赛扬

维普期刊数据库

Clustering by Fast Search and Find of Density Peaks with Data Field

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2016年第3期25卷 397-402页

作者： WANG Shuliang WANG Dakui LI Caoyuan LI Yan DING Gangyi International school of software Wuhan University school of software Beijing Institute of Technology

A clustering algorithm named "Clustering by fast search and find of density peaks" is for finding the centers of clusters quickly. Its accuracy excessively depended on the threshold, and no efficient way was given to select its suitable value, i.e., the value was suggested be estimated on the basis of empirical experience. A new way is proposed to automatically extract the optimal value of threshold by using the potential entropy of data field from the original dataset. For any dataset to be clustered, the threshold can be calculated from the dataset objectively instead of empirical estimation. The results of comparative experiments have shown the algorithm with the threshold from data field can get better clustering results than with the threshold from empirical experience.

关键词： Data field Potential entropy Big data clustering Optimal threshold value Automatic extraction

HTDcr:a job execution framework for high-throughput computing on supercomputers

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第1期67卷 69-85页

作者： Jiazhi JIANG Dan HUANG Hu CHEN Yutong LU Xiangke LIAO school of computer Science and Engineering Sun Yat-sen University school of software Engineering South China University of Technology

High-throughput computing(HTC) is a computing paradigm that aims to accomplish jobs by easily breaking them into smaller,independent ***,it requires a large amount of computing power for a long *** existing HTC frameworks are job-oriented without support for coscheduling with hardware architecture and task-level ***,most of the frameworks reach a limited scale,and their usability needs further ***,we present HTDcr,a job execution framework for the HTC on *** study aims to improve the throughput,task dispatching,and usability of the *** detail,the throughput optimizations include a sophisticated designed task management system,a hierarchical scheduler,and the co-optimization of the task-scheduling strategy with the application and hardware *** optimizations for usability include a programable execution workflow,mechanisms for more robust and reliable service qualities,and a fine-grained resource allocation system for the colocation of multiple *** to our evaluations,HTDcr can achieve outstanding scalability and high throughput on large-scale clusters for the HTC *** evaluate HTDcr with several microbenchmarks and real-world applications on Tianhe-2 and Sunway TaihuLight to demonstrate its effects on existing design *** instance,the task scheduling for two real-world applications integrated with the application and hardware characteristics achieves 1.7× and 1.9× speedups over the basic task-scheduling strategy.

关键词： high-throughput computing supercomputer task scheduling middleware password guessing

同方期刊数据库

Adversarial Training for Supervised Relation Extraction

在线全文

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2022年第3期27卷 610-618页

作者： Yanhua Yu Kanghao He Jie Li school of computer Science(National Pilot software Engineering school) Beijing University of Posts and TelecommunicationsBeijing 100876China

Most supervised methods for relation extraction(RE) involve time-consuming human annotation. Distant supervision for RE is an efficient method to obtain large corpora that contains thousands of instances and various relations. However, the existing approaches rely heavily on knowledge bases(e.g., Freebase), thereby introducing data noise. Various relations and noisy labeling instances make the issue difficult to solve. In this study, we propose a model based on a piecewise convolution neural network with adversarial training. Inspired by generative adversarial networks, we adopt a heuristic algorithm to identify noisy datasets and apply adversarial training to RE. Experiments on the extended dataset of SemEval-2010 Task 8 show that our model can obtain more accurate training data for RE and significantly outperforms several competitive baseline models. Our model has an F1 score of 89.61%.

关键词： relation extraction piecewise convolution neural network adversarial training generative adversarial network

维普期刊数据库

Steganalysis of Low Embedding Rates LSB Speech Based on Histogram Moments in Frequency Domain

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2017年第6期26卷 1254-1260页

作者： YANG Wanxia TANG Shanyu LI Miaoqi CHENG Yongfeng ZHOU Zhili school of computer Science China University of Geosciences College of Technology Gansu Agricultural University school of computer and software Nanjing University of Information Science and Technology

As wavelet packet transform is able to focus on minute change of signals, this study proposes an analytic approach of low embedding steganograpy based on high order Histogram moments in frequency domain(HMFD), which provides a key solution to the feature selection and extraction of HMFD. The detection results are tested with the LSB matching steganograpy of different embedding rates in speech signals, respectively, it is proved that the detection performance with HMFD applied is greater than that of histogram statistical *** by Wavelet packet decomposition(WPD) can effectively detect low embedding rates Least significant bit(LSB) speech steganography, its accuracy can be 60.8%while the embedding rate is only 3%.

关键词： Steganalysis Statistical moments of his togram Wavelet packet decomposition(WPD) Voice over IP(VoIP) Speech steganography

A Mixed Non-local Prior Model for Image Super-resolution Reconstruction

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2017年第4期26卷 778-783页

作者： ZHAO Shengrong LYU Zehua LIANG Hu Mudar SAREM school of computer Science and Technology Huazhong University of Science and Technology school of software Engineering Huazhong University of Science and Technology

Generating high-resolution image from a set of degraded low-resolution images is a challenge problem in image processing. Due to the ill-posed nature of Super-resolution(SR), it is necessary to find an effective image prior model to make it well-posed. For this purpose, we propose a mixed non-local prior model by adaptively combining the non-local total variation and non-local H1 models, and establish a multi-frame SR method based on this mixed non-local prior model. The unknown Highresolution(HR) image, motion parameters and hyperparameters related to the new prior model and noise statistics are determined automatically, resulting in an unsupervised super-resolution method. Extensive experiments demonstrate the effectiveness of the proposed SR method,which can not only preserve image details better but also suppress noise better.

关键词： Super-resolution(SR) Bayesian frame work Non-local H1 Non-local total variation Non-local edge & texture preserving(NLE&TP)

Enhancing Differential Evolution with Commensal Learning and Uniform Local Search

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2017年第4期26卷 725-733页

作者： PENG Hu WU Zhijian DENG Changshou State Key Laboratory of software Engineering School of Computer Wuhan University school of Information Science and Technology Jiujiang University

Differential evolution(DE) is a popular and powerful evolutionary algorithm for global optimization problems. However, the combination of mutation strategies and parameter settings of DE is problem dependent and choosing the suitable one is a challenge work and timeconsuming. In addition, the deficiency in local exploitation also has a significant influence on the performance of *** order to solve these problems, a DE variant with Commensal learning and uniform local search(CUDE) has been proposed in this paper. In CUDE, commensal learning is proposed to adaptively select optimal mutation strategy and parameter setting simultaneously under the same criteria. Moreover, uniform local search enhances exploitation ability. Comprehensive experiment results on all the CEC2013 test suite and comparison with the state-of-the-art DE variants indicate that the CUDE is very competitive.

关键词： Differential evolution Commensal learn ing Uniform local search Global optimization