检索结果-内蒙古大学图书馆

arXiv 2017年

作者： Du, Changde Du, Changying He, Huiguang Beijing China Laboratory of Parallel Software and Computational Science Institute of Software CAS Beijing China Center for Excellence in Brain Science and Intelligence Technology CAS Shanghai China

Decoding human brain activities via functional magnetic resonance imaging (fMRI) has gained increasing attention in recent years. While encouraging results have been reported in brain states classification tasks, reconstructing the details of human visual experience still remains difficult. Two main challenges that hinder the development of effective models are the perplexing fMRI measurement noise and the high dimensionality of limited data instances. Existing methods generally suffer from one or both of these issues and yield dissatisfactory results. In this paper, we tackle this problem by casting the reconstruction of visual stimulus as the Bayesian inference of missing view in a multiview latent variable model. Sharing a common latent representation, our joint generative model of external stimulus and brain response is not only "deep" in extracting nonlinear features from visual images, but also powerful in capturing correlations among voxel activities of fMRI recordings. The nonlinearity and deep structure endow our model with strong representation ability, while the correlations of voxel activities are critical for suppressing noise and improving prediction. We devise an efficient variational Bayesian method to infer the latent variables and the model parameters. To further improve the reconstruction accuracy, the latent representations of testing instances are enforced to be close to that of their neighbours from the training set via posterior regularization. Experiments on three fMRI recording datasets demonstrate that our approach can more accurately reconstruct visual stimuli. Copyright © 2017, The Authors. All rights reserved.

关键词： Brain

来源：评论

学校读者我要写书评

暂无评论

Semi-supervised bayesian deep multi-modal emotion recognition

arXiv

引用

arXiv 2017年

作者： Du, Changde Du, Changying Li, Jinpeng Zheng, Wei-Long Lu, Bao-Liang He, Huiguang Beijing China Laboratory of Parallel Software and Computational Science Institute of Software Cas Beijing Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China

In emotion recognition, it is difficult to recognize human's emotional states using just a single modality. Besides, the annotation of physiological emotional data is particularly expensive. These two aspects make the building of effective emotion recognition model challenging. In this paper, we first build a multi-view deep generative model to simulate the generative process of multi-modality emotional data. By imposing a mixture of Gaussians assumption on the posterior approximation of the latent variables, our model can learn the shared deep representation from multiple modalities. To solve the labeled-data-scarcity problem, we further extend our multi-view model to semi-supervised learning scenario by casting the semi-supervised classification problem as a specialized missing data imputation task. Our semi-supervised multi-view deep generative framework can leverage both labeled and unlabeled data from multiple modalities, where the weight factor for each modality can be learned automatically. Compared with previous emotion recognition methods, our method is more robust and flexible. The experiments conducted on two real multi-modal emotion datasets have demonstrated the superiority of our framework over a number of competitors. Copyright © 2017, The Authors. All rights reserved.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Optimizing Sparse Matrix Vector Multiplication Using Diagonal Storage Matrix Format

Optimizing Sparse Matrix Vector Multiplication Using Diagona...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Liang Yuan Yunquan Zhang Xiangzheng Sun Ting Wang Chinese Academy of Sciences China State Key Laboratory of Computer Science Chinese Academy and Sciences China Laboratory of parallel software and Computational Science IS-CAS China

Sparse matrix vector multiplication (SpMV) is used in many scientific computations. The main bottleneck of this algorithm is memory bandwidth and many methods reduce memory bandwidth usage by compressing the index array. The matrices from finite difference modeling applications often have several dense diagonals and sparse diagonals. For these matrices, the index array can be deleted by using diagonal storage format (DIA) to store dense diagonals and DIA & CSR mixed algorithm. In this paper we propose two improved sparse matrix storage format based on DIA format and the corresponding SpMV algorithms. We present the performance results on two platforms, which show that our method can reduce the memory usage for a wide range of sparse matrices and achieve speedup up to 1.87.

关键词： Arrays Sparse matrices Indexes Bandwidth Finite element methods Registers Algorithm design and analysis

来源：评论

学校读者我要写书评

暂无评论

Online Variational Bayesian Support Vector Regression

Online Variational Bayesian Support Vector Regression

引用

International Joint Conference on Neural Networks

作者： Siqi Deng Kan Gao Changying Du Wenjing Ma Guoping Long Yucheng Li Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing 100190 China Information Center of China Association for Science and Technology Beijing 100863 China

ISBN: (纸本)9781509006212

Traditional Support Vector Regression (SVR) solvers require user pre-specified penalty (regularization) parameter as input and typically model the training data with maximum a posterior (MAP) principle. The resultant point estimates can be affected seriously by inappropriate regularization, outliers and noise, especially when training online. In this paper, we address the aforementioned problems by developing a Bayesian SVR model with the pseudo-likelihood and data augmentation idea. Then we perform variational posterior inference in an augmented variable space and the approximate posterior of model weights, rather than point estimates as in traditional SVR, are used to make robust predictions. Besides, once the approximate posterior is obtained from a given set of data, we can regard it as model prior when dealing with new arrival data, which leads to a natural way to extend our batch model to the online scenario. Experiments on several benchmark regression problems as well as a real vehicle accident rate prediction task show that our models have superior performance while inferring penalty parameter automatically.

关键词： Bayes methods Support vector machines Data models Training Predictive models computational modeling Adaptation models

来源：评论

学校读者我要写书评

暂无评论

A new probabilistic algorithm for approximate model counting

arXiv

引用

arXiv 2017年

作者： Ge, Cunjing Ma, Feifei Liu, Tian Zhang, Jian State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences University of Chinese Academy of Sciences School of Electronics Engineering and Computer Science Peking University

Constrained counting is important in domains ranging from artificial intelligence to software analysis. There are already a few approaches for counting models over various types of constraints. Recently, hashing-based approaches achieve both theoretical guarantees and scalability, but still rely on solution enumeration. In this paper, a new probabilistic polynomial time approximate model counter is proposed, which is also a hashing-based universal framework, but with only satisfiability queries. A variant with a dynamic stopping criterion is also presented. Empirical evaluation over benchmarks on propositional logic formulas and SMT(BV) formulas shows that the approach is promising. Copyright © 2017, The Authors. All rights reserved.

关键词： Formal logic

来源：评论

学校读者我要写书评

暂无评论

A novel workflow engine method and metadata directory method to meet the mapping requirements between data and services

引用

International Review on Computers and software 2012年第7期7卷 3589-3593页

作者： Zhang, Jianwei Teng, Yuntian Deng, Pan Cai, Ming Ma, Shilong State Key Laboratory of Software Development Environment BeiHang University Beijing China Institute of Geophysics China Earthquake Administration Beijing China Lab of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing China

National Earthquake Data Sharing and Exchanging Platform (i.e. National Earthquake Data Grid) is used to solve the demands of data sharing, interoperability and coordinately access in earthquake industry. However, the heterogeneous, distributed and variety data resources make it difficult to share and coordinately access. This paper first propose a novel workflow engine method, which uses Globus service to package heterogeneous resources to achieve the unified access, workflow technology to achieve the coordinately access in distributed environment, and metadata directory method to meet the mapping requirements between data and services. Then we implement this workflow engine in Earthquake Data Grid. Experiment results show that our system can be expanded and applied to other appropriate industries. © 2012 Praise Worthy Prize S.r.l.

关键词： Earthquakes

来源：评论

学校读者我要写书评

暂无评论

Resource Scheduling Strategies for Cloud Editors for Low-Code Development

Resource Scheduling Strategies for Cloud Editors for Low-Cod...

引用

Electronics and Communication, Network and Computer Technology (ECNCT), International Conference on

作者： Yuhao Gao Shufeng Duan Wen Bo Changyou Zhang School of Information Science and Technology Shijiazhuang Tiedao University Shijiazhuang China Laboratory of Parallel Software and Computational Science University of Chinese Academy of Sciences Beijing China

software ecosystems are accelerating the vertical division of labour and horizontal integration in the software industry, attracting widespread attention from industry and academia. The low-code development cloud editor oriented in this paper is a generalisation of building software ecosystems, which reduces the technical threshold of developers in the ecosystem and improves the development efficiency by supplementing the interfaces and services and customising the cloud editor platform to support core technologies such as multi-language, sharing, service publishing, and low-code. In this context, this paper focuses on the editor resource scheduling strategy for cloud editor: 1) for low-code programming, a collaborative service planning method is proposed to start containers running synchronously on demand in units of service components, avoiding the start of temporary unnecessary service startups and improving the resource utilisation rate; 2) designing a threetier platform service framework that supports on-demand running of component code, and formulating a related service strategy that Improve the efficiency of component execution; 3) For three different scenarios, a resource scheduling method based on greedy algorithm is proposed to reasonably utilise resources; 4) Experiments verify the effect of the algorithmic strategy in this paper in terms of resource scheduling efficiency, quality and stability. The work in this paper is used in project development practice to support low-code development software ecosystem.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Approach for Malware Detection by Using APKs

Machine Learning Approach for Malware Detection by Using APK...

引用

2017 2nd International Conference on Computer, Network Security and Communication Engineering(CNSCE2017)

作者： Rubata RIASAT Muntaha SAKEENA Abdul Hannan SADIQ Chong WANG Chang-you ZHANG Yong-ji WANG State key laboratory of Computer Science Institute of Software Chinese Academy of Sciences University of Chinese Academy of Sciences College of Electrical and Mechanical Engineering National University of Science and Technology Comsats Institute of Technology Laboratory of parallel Software and Computational Science Institute of SoftwareChinese Academy of Sciences

An android mobile phone is one of the most anticipated smart phone operating systems on the market. The open source Android platform not only allow developer to take full advantage of the mobile system, but also raises significant issues related to malicious applications(Apps). Although understanding the android malware using dynamic analysis can provide a compressive view and it is still subjected to high cost in environment development and manual effort in investigation. In this study our proposed approach provides a static and dynamic analyst paradigm for detecting android malware. The mechanism considers the static information including permissions, deployment of components, intent message passing and API calls for characterizing the android application behavior. In order to recognize different intentions of android malware, different kind of clustering and classification can be applied to enhance the malware detection capability. Our approach extracts the information from the each APKs manifest file, and regards component(Activity, Receiver, Services) as entry points drilling down for tracing API calls related to permissions. Next it applies detection methods based on matching(Signature matching) and learning(SVM and random forest) algorithm to classify the applications as benign or malicious. The experimental results show that the accuracy of our approach is better than one of the well-known tool, Androguard, also it is efficient since its take half of the time than Androguard to predict 600 applications as benign or malicious. The open source Android platform allows developer to take full advantage of the mobile.

关键词： Machine learning Malware detection Android Reverse engineering APKs

来源：评论

学校读者我要写书评

暂无评论

CLSIFT: An Optimization Study of the Scale Invariance Feature Transform on GPUs

CLSIFT: An Optimization Study of the Scale Invariance Featur...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Weiyan Wang Yunquan Zhang Long Guoping Shengen Yan Haipeng Jia School of Computer and Control Engineering University of Chinese Academy of Sciences Beijing China State Key Laboratory of Computing Science Institute of Software Chinese Academy of Sciences Lab. of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences

Scale Invariance Feature Transform (SIFT) is quite suitable for image matching because of its invariance to image scaling, rotation and slight changes in illumination or viewpoint. However, due to high computation complexity it's technically challenging to deploy SIFT in real time application situations. To address this problem, we propose CLSIFT, an OpenCL based highly speeded up and performance portable SIFT solution. Important optimization techniques employed in CLSIFT such as: (1) For less global memory traffic, independent logical functions are merged into the same kernel to reuse data. (2) loop buffers are introduced in for data and intermediate results reusing. (3)Task queue used to schedule threads in the same branch to remove branch divergences. (4) Data partition is based on the statics patterns for workload balance among workgroups. (5) Overlap of CPU time and better parallel strategies are used too. With all mentioned efforts, CLSIFT processes lena. jpg at 74.2 FPS and 43.4FPS respectively on NVidia and AMD GPUS, much higher than CPU's nearly 10 FPS and the known fastest SIFTGPU's 39.8 FPS and 13FPS. Moreover in a quantitative comparison approach we analyze those successful strategies beating SIFTGPU, a famous existing GPU implementation. Additionally, we observe and conclude that NVidia GPU achieves better occupancy and performance due to some factors. Finally, we summarize some techniques and empirical guiding principles that may be shared by other applications on GPU.

关键词： Graphics processing units Instruction sets Kernel Accuracy Histograms Optimization Memory management

来源：评论

学校读者我要写书评

暂无评论

Can graph neural networks learn to solve MaxSAT problem?

arXiv

引用

arXiv 2021年

作者： Liu, Minghao Jia, Fuqi Huang, Pei Zhang, Fan Sun, Yuchen Cai, Shaowei Ma, Feifei Zhang, Jian State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Inspir.ai Beijing China Beijing China

With the rapid development of deep learning techniques, various recent work has tried to apply graph neural networks (GNNs) to solve NP-hard problems such as Boolean Satisfiability (SAT), which shows the potential in bridging the gap between machine learning and symbolic reasoning. However, the quality of solutions predicted by GNNs has not been well investigated in the literature. In this paper, we study the capability of GNNs in learning to solve Maximum Satisfiability (MaxSAT) problem, both from theoretical and practical perspectives. We build two kinds of GNN models to learn the solution of MaxSAT instances from benchmarks, and show that GNNs have attractive potential to solve MaxSAT problem through experimental evaluation. We also present a theoretical explanation of the effect that GNNs can learn to solve MaxSAT problem to some extent for the first time, based on the algorithmic alignment theory. © 2021, CC BY.

关键词： Deep learning

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：