检索结果-内蒙古大学图书馆

International conference on parallel and distributed Systems (ICPADS)

作者： Canhui Luo Xuan Liu Xinning Chen Juan Luo College of Computer Science and Electronic Engineering Hunan University P. R. China The Science and Technology on Parallel and Distributed Processing Laboratory (PDL) P. R. China

ISBN: (数字)9781728190747

ISBN: (纸本)9781728183824

Multi-agent reinforcement learning system is used to solve the problem that agents achieve specific goals in the interaction with the environment through learning policies. Almost all existing multi-agent reinforcement learning methods assume that the observation of the agents is accurate during the training process. It does not take into account that the observation may be wrong due to the complexity of the actual environment or the existence of dishonest agents, which will make the agent training difficult to succeed. In this paper, considering the limitations of the traditional multi-agent algorithm framework in noisy environments, we propose a multi-agent fault-tolerant reinforcement learning (MAFTRL) algorithm. Our main idea is to establish the agent's own error detection mechanism and design the information communication medium between agents. The error detection mechanism is based on the autoencoder, which calculates the credibility of each agent's observation and effectively reduces the environmental noise. The communication medium based on the attention mechanism can significantly improve the ability of agents to extract effective information. Experimental results show that our approach accurately detects the error observation of the agent, which has good performance and strong robustness in both the traditional reliable environment and the noisy environment. Moreover, MAFTRL significantly outperforms the traditional methods in the noisy environment.

关键词： Training Fault tolerance Working environment noise Fault tolerant systems Reinforcement learning Robustness Noise measurement

来源：评论

学校读者我要写书评

暂无评论

Short-Term Rainfall Estimation by Machine Learning methods 15

Short-Term Rainfall Estimation by Machine Learning Methods

引用

IEEE 15th International Scientific conference on Informatics (INFORMATICS)

作者： Krammer, Peter Kvassay, Marcel Habala, Ondrej Hluchy, Ladislav Slovak Acad Sci Dept Parallel & Distributed Informat Proc Inst Informat Bratislava Slovakia

ISBN: (纸本)9781728131795

Modeling stochastic systems is a real challenge in many areas. Even the meteorology domain is not an exception;the modeling of precipitation activity is markedly stochastic and is influenced by a number of related physical variables (temperature, pressure, humidity, wind). Accurate precipitation estimation is thus highly non-trivial. Today's technical capability, automated data measuring (whether using Radar or automatic meteorological stations), as well as subsequent large-scale data processing and regression model training, allow the meteorological estimations and predictions with increasing accuracy. This paper demonstrates selected uses of artificial neural networks in the field of meteorology, as well as solving problems with pre-processing and integrating time-spatial meteorological data.

关键词： Neural network Machine learning Data mining Modeling Meteorology Rainfall estimation Data integration Data cleaning Time-spatial data

来源：评论

学校读者我要写书评

暂无评论

A Comparison of Random Task Graph Generation methods for Scheduling Problems 25th

A Comparison of Random Task Graph Generation Methods for Sch...

引用

25th International conference on parallel and distributed Computing (Euro-Par)

作者： Canon, Louis-Claude El Sayah, Mohamad Heam, Pierre-Cyrille Univ Bourgogne Franche Comte CNRS FEMTO ST Inst Besancon France

ISBN: (纸本)9783030294007;9783030293994

How to generate instances with relevant properties and without bias remains an open problem of critical importance to compare heuristics fairly. When scheduling with precedence constraints, the instance is a task graph that determines a partial order on task executions. To avoid selecting instances among a set populated mainly with trivial ones, we rely on properties such as the mass, which measures how much a task graph can be decomposed into smaller ones. This property and an in-depth analysis of existing random instance generators establish the sub-exponential generic time complexity of the studied problem.

关键词： Scheduling

来源：评论

学校读者我要写书评

暂无评论

Data-parallel computational model for next generation sequencing on commodity clusters 1

引用

15th International conference on parallel Computing Technologies, PaCT 2019

作者： Hajibaba, Majid Sharifi, Mohsen Gorgin, Saeid Department of Electrical and Information Technology Iranian Research Organization for Science and Technology Tehran Iran Department of Computer Engineering Iran University of Science and Technology Tehran Iran

ISBN: (数字)9783030256364

ISBN: (纸本)9783030256357

It is obvious that the next generation sequencing (NGS) technologies, are poised to be the next big revolution in personalized healthcare, and caused the amount of available sequencing data growing exponentially. While NGS data processing has become a major challenge for individual genomic research, commodity computers as a cost-effective platform for distributed and parallel processing in laboratories can help processing such huge volume of data. To deploy sequence-processing methods on these platforms, in this paper we present a parallel computational model for BLAST on commodity clusters that works in a data parallel manner. The suggested model has a master-worker paradigm. The master stores temporarily incoming requests and splits the database to chunks according to the number of available workers. Each worker pulls, formats, and searches queries against a unique chunk of the database. To show that our model works well, we used queries with different lengths to search against a small database (i.e. UniProtKB/SWISS-PROT) and a large database (i.e. UniProtKB/TrEMBL). The results were equal with the output of the golden method (i.e. NCBI BLAST) and the performance of our model outperformed the most popular distributed form of BLAST (i.e. mpiBLAST) with 25% higher performance. © 2019, Springer Nature Switzerland AG.

关键词： Cost effectiveness

来源：评论

学校读者我要写书评

暂无评论

Transfer Learning for Airport Detection Based on Saliency Fusion of parallel Lines and Regions of Interest 7

Transfer Learning for Airport Detection Based on Saliency Fu...

引用

7th IEEE International conference on Computer Science and Network Technology (ICCSNT)

作者： Huang, Weihua Xia, Jing Xu, Mingdi Huang, Jinzhou Wuhan Digital Engn Inst AI Res Ctr Wuhan Peoples R China

ISBN: (纸本)9781728132990

Most existing airport detection methods for remote sensing image utilizes linear features of the airport runway insufficiently, and the computational complexity is high due to multi-scale anchor matching and global searching within full image. To solve this problem, an airport detection method based on saliency fusion of parallel lines and regions of interest is presented in this paper. Firstly the parallelism feature of the airport runway is extracted based on the prior knowledge of airport, and then regions of interest (ROI) is obtained according to the improved graph based visual saliency (GBVS). The airport is then located through the saliency fusion of parallel lines and regions of interest. Finally airport detection is achieved by transfer learning. The experimental results demonstrate that the proposed method is more advantageous than the comparison algorithms in terms of detection accuracy, processing speed, and false alarm rate. Moreover, the method only requires a small amount of samples for model training.

关键词： airport detection parallel line graph based visual saliency saliency fusion transfer learning

来源：评论

学校读者我要写书评

暂无评论

Integrating Long-Short Term Network for Efficient Video Object Segmentation 31

Integrating Long-Short Term Network for Efficient Video Obje...

引用

31st British Machine Vision conference, BMVC 2020

作者： Wang, Jingjing Teng, Zhu Zhang, Baopeng Fan, Jianping School of Computer and Information Technology Beijing Jiaotong University Beijing China AI Lab Lenovo Research Beijing China

Real-world application of video object segmentation (VOS) is a very challenging problem, especially for multiple video object segmentation. The deep-learning-based approaches have recently dominated VOS by fine-tuning the networks at the first frame to seize the object dynamics, but they may result in impractical frame-rates and risk of over-fitting. To overcome this limitation, we develop an efficient and fully end-to-end model to achieve fast and accurate VOS, named Long-Short Term Network (LSTNet). It contains a long term network to encode absolute object variations and a short term network to capture relative object dynamics. The segmentation results of video objects can be directly acquired by an attentional gate operation based on these two networks. Our proposed model runs at a very high speed and can conveniently tackle multi-object segmentation without post-processing. Extensive experiments on widely used benchmarks including YouTube-VOS and DAVIS 2017 have demonstrated that our proposed model can achieve a competitive accuracy and speed in comparison to a number of state-of-the-art methods. © 2020. The copyright of this document resides with its authors. It may be distributed unchanged freely in print or electronic forms.

关键词： image segmentation

来源：评论

学校读者我要写书评

暂无评论

Approximate Nearest-Neighbour Fields via Massively-parallel Propagation-Assisted K-D Trees

Approximate Nearest-Neighbour Fields via Massively-Parallel ...

引用

IEEE International conference on Big Data

作者： Cosmin Eugen Oancea Ties Robroek Fabian Gieseke University of Copenhagen Copenhagen Denmark University of Münster Münster Germany

ISBN: (数字)9781728162515

ISBN: (纸本)9781728162522

Nearest neighbour fields accurately and intuitively describe the transformation between two images and have been heavily used in computer vision. Generating such fields, however, is not an easy task due to the induced computational complexity, which quickly grows with the sizes of the images. Modern parallel devices such as graphics processing units depict a viable way of reducing the practical run time of such compute-intensive tasks. In this work, we propose a novel parallel implementation for one of the state-of-the-art methods for the computation of nearest neighbour fields, called p ropagation-assisted k -d trees. The resulting implementation yields valuable computational savings over a corresponding multi-core implementation. Additionally, it is tuned to consume only little additional memory and is, hence, capable of dealing with high-resolution image data, which is vital as image quality standards keep rising.

关键词： Computer vision Satellites Memory management Big Data Task analysis Standards Monitoring

来源：评论

学校读者我要写书评

暂无评论

Improving Robustness of DNNs against Common Corruptions via Gaussian Adversarial Training

Improving Robustness of DNNs against Common Corruptions via ...

引用

IEEE Visual Communications and image processing (VCIP)

作者： Chenyu Yi Haoliang Li Renjie Wan Alex C. Kot School of Electrical and Electronic Engineering Nanyang Technological University Singapore

ISBN: (数字)9781728180687

ISBN: (纸本)9781728180694

Deep neural networks have demonstrated tremendous success in image classification, but their performance sharply degrades when evaluated on slightly different test data (e.g., data with corruptions). To address these issues, we propose a minimax approach to improve common corruption robustness of deep neural networks via Gaussian Adversarial Training. To be specific, we propose to train neural networks with adversarial examples where the perturbations are Gaussian-distributed. Our experiments show that our proposed GAT can improve neural networks' robustness to noise corruptions more than other baseline methods. It also outperforms the state-of-the-art method in improving the overall robustness to common corruptions.

关键词： Robustness Perturbation methods Training Neural networks Standards Gaussian noise Tensors

来源：评论

学校读者我要写书评

暂无评论

REK-SVM: A Robust and Efficient SVM Algorithm Based on K-Medians Clustering 10

REK-SVM: A Robust and Efficient SVM Algorithm Based on K-Med...

引用

10th International conference on Graphics and image processing (ICGIP)

作者： Gao, Chongjun Sang, Nong Lei, Jiahui Huazhong Univ Sci & Technol Sch Automat Key Lab Minist Educ Image Proc & Intelligent Control Wuhan Hubei Peoples R China

ISBN: (数字)9781510628298

ISBN: (纸本)9781510628298

Support vector machines (SVMs) have been widely used for binary classification. But large-scale training set will bring huge computation to the SVM. Researcher have proposed many techniques to improve the training efficiency of SVMs, and a typical class of improved SVMs is based on sparsely reducing training samples. To achieve this, clustering-based methods are most commonly used. However, clustering-based methods are ready to be disturbed by noise points. In order to solve this problem, this paper proposes a robust and efficient SVM algorithm based on K-Medians clustering (REK-SVM). Here, for each cluster, the cluster center takes the median value of each dimension attribute in the cluster, which can reduce the noise points. Especially, when the number of noise points distributed discretely is less than half of the total number of samples in the cluster, noise interference can be completely removed. The noise-free or noise-reduced subset data is used to train the SVM model. Experimental results show that our algorithm is fast and effective. For the processing of noise-containing classification data, its performance far exceeds SVM in terms of classification accuracy and efficiency. Compared to the K-SVM, they have the same computational complexity, but our algorithm is much higher than K-SVM in classification accuracy.

关键词： SVM clustering denoising K-Medians

来源：评论

学校读者我要写书评

暂无评论

CACAI: Comparative analysis of crypto apps in India

引用

AIP conference Proceedings 2023年第1期2916卷

作者： Rahul Johari Anuj Talwar Suhas Harbola Jyotsna Yadav Himesh Nayak Deo Prakash Vidyarthi SWINGER: Security Wireless IoT Network Group of Engineering and Research University School of Information Communication and Technology (USICT) Guru Gobind Singh Indraprastha University Sector-16C Dwarka Delhi India Computer vision and Image processing lab University School of Information Communication and Technology (USICT) Guru Gobind Singh Indraprastha University Sector-16C Dwarka Delhi India School of Computer and Systems Sciences Parallel and Distributed Systems Lab JNU Delhi India

Crypto-currency is a decentralized digital currency which can be used as a medium of exchange. As it does not have any central authority, it is unregulated and a volatile asset. Globally the crypto market is worth a few trillion dollars. The Indian crypto market suffered a lot due to a ban imposed by the Reserve Bank of India (RBI), an Indian regulatory agency, which lasted for around two years. Even after the ban and cautions issued by the RBI, the crypto market in India touched a billion dollar mark in no time. Currently, the crypto market has more than 10 million Indian investors which is growing fast as the number of exchange apps in the Indian market increase with the ease of trading in crypto. In the current research work Comparative Analysis of Crypto App in India (CACAI) has been performed and results exceed the expectations.

关键词：

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：