2017 IEEE International Conference on Computer and Information Technology (CIT)最新文献_第2页

Learning From Big Data: A Survey and Evaluation of Approximation Technologies for Large-Scale Reinforcement Learning 从大数据中学习:大规模强化学习的近似技术综述与评价

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.11

Cheng Wu, Yiming Wang

{"title":"Learning From Big Data: A Survey and Evaluation of Approximation Technologies for Large-Scale Reinforcement Learning","authors":"Cheng Wu, Yiming Wang","doi":"10.1109/CIT.2017.11","DOIUrl":"https://doi.org/10.1109/CIT.2017.11","url":null,"abstract":"A key problem in large-scale reinforcement learning is to deal with big data, in terms of a very large number of environment states and many possible actions. Function approximation can improve the ability of a reinforcement learner to solve large-scale problems. Tile coding and Kanerva coding are two classical methods for implementing function approximation, but these methods may give poor performance when applied to large-scale, high-dimensional instances. In the paper, we evaluate a collection of hard instances of the predator-prey pursuit problem, a classic reinforcement learning platform with scalable state-action space, to compare these two methods and their optimization techniques. We first show that Kanerva coding gives better results than Tile coding when the dimension of the instances increases. We then describe a feature optimization mechanism and show that it can increase the fraction of instances that are solved by both Tile coding and Kanerva coding. Finally, we demonstrate that a fuzzy approach to function approximation can further increase the fraction of instances. We show that our fuzzy approach to Kanerva coding outperforms fuzzy Tile coding when feature optimization is applied. We conclude that discrete and fuzzy Kanerva coding represent powerful function approximation techniques that can outperform discrete and fuzzy Tile coding on large-scale, high-dimensional learning problems.","PeriodicalId":378423,"journal":{"name":"2017 IEEE International Conference on Computer and Information Technology (CIT)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130618618","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

A New Method of Error Reduction for Affine Projection Algorithm 一种新的仿射投影算法误差减小方法

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.18

S. Koike

引用次数: 0

Software-Defined Edge Defense Against IoT-Based DDoS 针对物联网DDoS的软件定义边缘防御

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.61

Ihsan Mert Ozcelik, Niaz Chalabianloo, Gürkan Gür

引用次数: 73

3D Matrix-Based Visualization System of Association Rules 基于三维矩阵的关联规则可视化系统

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.39

Tingting Zhang, Zheng Chang, T. Ristaniemi, Guohua Liu

引用次数: 4

Cloud IEC 61850: DDS Performance in Virtualized Environment with OpenDDS 云IEC 61850:使用OpenDDS的虚拟化环境中的DDS性能

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.17

R. Ferreira, R. S. Oliveira

{"title":"Cloud IEC 61850: DDS Performance in Virtualized Environment with OpenDDS","authors":"R. Ferreira, R. S. Oliveira","doi":"10.1109/CIT.2017.17","DOIUrl":"https://doi.org/10.1109/CIT.2017.17","url":null,"abstract":"The idea of using Virtualization and Cloud as the underlying infrastructure for future automation systems in the context of the Internet of Things has been presented in the literature. This paper presents the assessment of the end-to-end performance of a distributed and virtualized architecture from the DDS communication middleware layer, running the performance test benchmark applications of the OpenDDS implementation. The experimental distributed and virtualized environment used considers two ordinary physical hosts connected by a single network. The virtualization layer is realized with KVM as hosted hypervisor with Ubuntu Linux and TinyCore as Guest-OS. A set of test scenarios combining host-host, host-VM and VM-VM communications is defined. Each test consist of publishing circa 11 thousand messages of 250 bytes for each one of the different transport protocol supported by OpenDDS. The round-trip time is measured for each message or sample, from which the respective latency and jitter are calculated. The results obtained are presented in tabular and graphical formats. The results show that soft real-time applications with deadlines up to 1 ms could be run in such virtualized environment. We show that is reasonable to consider the potential of running automation applications on a distributed and virtualized environment. The contribution of this study is to provide information, based on the proposed reference architecture and its time characterization, to engineers responsible for the design of automation solutions to start considering and assessing the use of modern computing architectures and technologies, specially virtualization and communication middlewares, in new engineering projects.","PeriodicalId":378423,"journal":{"name":"2017 IEEE International Conference on Computer and Information Technology (CIT)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129217063","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 6

Dynamic Reconfiguration of Architecture in the Communication Network of Air Traffic Management System 空中交通管理系统通信网络体系结构的动态重构

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.13

I. Kabashkin

{"title":"Dynamic Reconfiguration of Architecture in the Communication Network of Air Traffic Management System","authors":"I. Kabashkin","doi":"10.1109/CIT.2017.13","DOIUrl":"https://doi.org/10.1109/CIT.2017.13","url":null,"abstract":"Communication between air traffic controllers and pilots remains a vital part of air traffic control operations, and communication problems can result in hazardous situations. The modern Air Traffic Management (ATM) system has independent direct communication channels (CC) for each controllers operating at different radio frequencies. Currently, the main method of improving the reliability of controller's CC is redundancy of equipment to provide communications on each frequency of interaction ground-to-air channel. The fault tolerance of such method is limited by number of redundancy elements. ATM communication network has periodical sessions of communications with relatively short active periods of communication interactions alternate with relatively long pauses in the translation of signals. The resilience of such network can be improved by dynamic architecture reconfiguration of communication equipment for operation in different channels during communication sessions. In paper the reliability of Air Traffic Management voice communication network with periodical sessions of communications for real conditions of ATM and with different strategies of reconfiguration for set of radio stations is investigated. Mathematical model of the channel reliability is developed. Comparative analysis of reliability for proposed strategies of ATM communication network reconfiguration is performed.","PeriodicalId":378423,"journal":{"name":"2017 IEEE International Conference on Computer and Information Technology (CIT)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126875008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

Investigating the Agility Bias in DNS Graph Mining DNS图挖掘中的敏捷性偏差研究

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.55

Jukka Ruohonen, V. Leppänen

引用次数: 4

A Simplified Massive MIMO Approach Using Precoding 一种简化的大规模MIMO预编码方法

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.12

M. Silva, R. Dinis

引用次数: 3

CDController: A Cloud Data Access Control System Based on Reputation CDController:基于信誉的云数据访问控制系统

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.20

Huaqing Lin, Zheng Yan, R. Kantola

引用次数: 0

Remarks on Adaptive-Type Hypercomplex-Valued Neural Network-Based Feedforward Feedback Controller 自适应型超复值神经网络前馈反馈控制器评述

2017 IEEE International Conference on Computer and Information Technology (CIT) Pub Date : 2017-08-01 DOI: 10.1109/CIT.2017.16

Kazuhiko Takahashi

引用次数: 11