5G时代的AI技术应用详解

5G时代的AI技术应用详解"

作者:亚信科技(中国)有限公司
ISBN:9787302565321
定价:¥79
字数:千字
页数:
出版时间:2020.11.01
开本:
版次:1-1
装帧:
出版社:清华大学出版社
简介

本书结合大量实际案例,全面且详细地介绍了企业在5G时代应该如何应用AI技术来提升

生产、运营和管理能力。全书共分为三篇:第一篇为基础与网络篇,包括第1~4章,主要介绍

如何将AI技术应用于网络智能切片、物联网和5G网络多量纲计费业务场景中;第二篇为客户

与管理篇,包括第5~8章,以客户体验管理、客户关系管理、企业业务流程管理、企业商业智

能决策四大典型应用场景为例,详细介绍如何通过AI技术提升企业的管理效能;第三篇为运维

与安全篇,包括第9~12章,其中第9~11章分别介绍AI技术应用于网络智能运维、机房智

慧管控、智能安防的应用案例,第12章则对AI能力平台化的建设、沉积等内容进行详细论述,

并给出AI平台建设的理念、功能设计和技术设计建议。

本书可供通信行业和其他行业的IT从业人员,以及科研人员、高校师生阅读和参考。

前言

丛书序

2019年6月6日,工信部正式向中国电信、中国移动、中国联通和中国广电四家企业发放了5G牌照。这意味着中国正式按下了5G商用的启动键。可以想见,在未来的若年干中,万众睢睢的5G将与人工智能、云计算、大数据、物联网等新技术一起,改变个人生活,催生行业变革,加速经济转型,推动社会发展,真正打造一个“万物智联”的多维世界。

5G将带来个人生活方式的迭代。更加畅快的通信体验、无处不在的AR/VR、智能安全的自动驾驶……这些都将因5G的到来而变成现实,给人类带来更加自由、丰富、健康的生活体验。

5G将带来行业的革新。受益于速率的提升、时延的改善、接入设备容量的增加,5G触发的革新将从通信行业溢出,数字化改造得以加速,新技术的加持日趋显著,新的商业模式不断涌现,产业的升级将让千行百业脱胎换骨。

5G将带来多维的跨越。C端消费与B端产业转型将共振共生。“4G改变生活,5G改变社会”,5G时代,普通消费者会因信息技术再一次升级而享受更多便捷,千行百业的数字化、智能化转型也会真正实现,两者互为表里,互

相助推,把整个社会的变革提升到新高度。

2019年是5G元年,也是亚信科技(中国)有限公司(简称亚信科技)上市后的第一个财年。作为国内领先的软件与服务提供商、云网一体管理服务提供商,亚信科技紧扣时代发展节拍,积极拥抱5G、云计算、大数据、人工智能、物联网等先进技术,与业界客户、合作伙伴共同建设5G+X的生态体系,为5G赋能千行百业、企业数字化转型、产业可持续发展积极做出贡献。

在过去的一年中,亚信科技继续深耕电信业务支撑软件与服务(Business Supporting System,BSS)的优势领域,为三大运营商的5G业务在中华大地全面商用提供了强有力的支撑。

亚信科技将能力延展到5G网络Operation Supporting System(OSS)领域,公司打造的5G网络智能化产品在三大运营商取得了多个商用局点的突破与落地实践,在帮助运营商优化5G网络环境、提升5G服务体验的同时,公司也迈出了拓展OSS领域的坚实一步。

亚信科技在数字化运营的Data-Driven Software as a Service(DSaaS)这一创新业务板块也取得了规模化突破。在金融、交通、能源、政府等多个领域,帮助行业客户打造“数智”能力,用大数据和人工智能技术,协助他们获客、活客、留客,改善服务质量,实现行业运营数字化转型。

亚信科技在垂直行业市场服务领域进一步拓展,行业大客户版图进一步扩大,公司与云计算的各头部企业达成云MSP合作,持续提升云集成、云SaaS、云运营能力,并与他们一起,帮助邮政、能源、交通、金融、零售等数十个大型行业客户上云、用云,降低信息化支出,提升数字化效率。

亚信科技同时积极强化、完善了技术创新与研发的体系和机制。在过去的一年中,多项关键技术与产品获得了国际和国家级奖项,诸多技术组合形成了国际与国家标准。5G+ABCDT的灵动组合,重塑了包括亚信科技自身在内的行业技术生态体系。“5G与AI技术大系”丛书是亚信科技在过去几年中,以匠心精神打造我国5G软件技术体系的创新成果与科研经验的总结。我们非常高兴能将这些阶段性成果以丛书的形式与行业伙伴们分享与交流。

我国经历了从2G落后、3G追随、4G同步,到5G领先的历程。在这个过程中,亚信科技从未缺席。在未来的5G时代,我们将继续坚持以技术创新为引领,与业界合作伙伴们共同努力,为提升我国5G科技和应用水平、为国家新基建和全行业数字化转型贡献力量。

高念书

 

2020年9月于北京

2019年6月6日,工信部正式向中国电信、中国移动、中国联通、中国广电发放了5G商用牌照,标志着我国开启了真正意义上的5G商用元年。而2020年年初,国家新基建宏观政策则更是将5G列为了需要重点发展的七大关键领域之首。

很显然,5G的快速发展将使企业面临全新的发展环境,新技术、新网络和新业务的变革,为各行业带来了全新的发展机遇和挑战。如何在5G时代更好地规划网络以支撑业务应用?如何在降本增效的同时提升客户服务感知?如何科学合理地设计多量纲的计费模式?如何实现业务决策和业务流程的智能化。这些单纯依靠传统的方式已经无能为力,将AI技术应用于5G业务场景中、解决企业的运营管理问题,已成为大势所趋。但如何将AI技术应用到纷繁复杂的业务场景中,却是企业面临的一个非常实际的问题。

本书主要选取一些非常典型的5G业务场景,结合大量实际应用案例,系统介绍如何将AI与5G技术相结合,应用于5G智能网络、5G切片、物联网、Massive MIMO、CEM(客户体验管理)、CRM(客户关系管理)、5G智能多量纲计费、商业智能分析、智能安防等5G业务场景中。希望通过理解AI技术在这些业务场景中的应用,为企业在5G时代通过AI技术赋能生产和运营管理提供有益的参考。同时,在本书的最后一章,特别介绍了在5G时代如何通过平台建设的方式,将AI能力沉淀下来,更好地支持企业的业务管理运营。

全书共分为三篇:第一篇为基础与网络篇,包括第1~4章,主要介绍如何将AI技术应用于网络智能切片、物联网和5G网络多量纲计费业务场景中;第二篇为客户与管理篇,包括第5~8章,以客户体验管理、客户关系管理、企业业务流程管理、企业商业智能决策四大典型应用场景为例,详细介绍如何通过AI技术提升企业的管理效能;第三篇为运维与安全篇,包括第9~12章,其中第9~11章分别介绍AI技术应用于设备智能运维、机房智慧管控、智能安防的应用案例,第12章则对AI能力平台化的建设、沉积等内容进行详细论述,并给出AI平台建设的理念、功能设计和技术设计建议。

与市场上介绍5G和AI的其他图书相比,本书具有如下特色。

(1)定位明确。本书定位于AI技术与应用的有机结合,在应用中详细分析技术。本书非常透彻地分析了企业在5G时代面临的市场环境、发展机遇与挑战,并通过大量的技术细节,详细论述了企业在面对这些机遇和挑战时,应如何通过技术手段注智业务场景,并辅以详细的应用案例。

(2)技术先进。本书各章节中涉及的技术均为目前国内外AI领域的前沿技术,同时参考了亚信科技(中国)有限公司发表于国际权威学术刊物上的多篇学术论文,读者通过阅读本书,可以对5G时代AI前沿技术的应用有更加充分的了解。

(3)注重实战。本书中的技术应用实例大部分是亚信科技(中国)有限公司近年来在生产过程中积累的实际应用案例,实践性强。

本书由亚信科技通信人工智能实验室编写,编写组成员包括欧阳晔博士、孟祥德、白世明、杨爱东博士、薛明博士、曾树明博士、经琴、蒋炜、李红霞、宋勇、龚福才等。

由于编者水平有限,更兼时间和精力所限,书中不足之处在所难免,若蒙读者诸君不吝告知,必将不胜感激。

编者

2020年9月于北京

目录

第一篇   基础与网络篇

第1章  “5G+AI”概述·································2

1.1  新基建下的“5G+AI”技术发展·························3

1.1.1  新基建的内涵和外延···············································3

1.1.2  新基建对5G和AI发展的影响······························6

1.2 5G时代的AI技术趋势······································10

1.2.1 AI部署云边协同····················································10

1.2.2 AI注智实时持续····················································12

1.2.3 AI应用民主灵活····················································13

1.2.4 AI决策高度仿真····················································14

1.3  我国5G产业与技术发展···································16

1.3.1  我国5G技术发展历程··········································16

1.3.2 5G改变社会···························································17

1.4  我国AI产业与技术发展····································22

1.4.1  人工智能发展概述·················································22

1.4.2  我国人工智能技术的发展·····································24

第2章  AI与5G网络智能切片····················29

2.1 5G业务多样化与网络需求弹性化····················29

2.2 5G网络智能切片概述········································31

2.2.1 5G网络智能切片的概念与特征···························32

2.2.2 5G网络智能切片端到端结构·······························33

2.2.3 5G网络智能切片的RAN侧技术挑战················34

2.2.4 5G网络智能切片的AI平台和分析系统·············35

2.2.5 5G网络智能切片的智能部署·······························36

2.2.6 5G网络智能切片的标准化增强···························37

2.3  应用于5G网络切片中的AI技术·····················38

2.3.1 5G网络智能切片的设计流程·······························38

2.3.2  基于GA-PSO优化的网络切片编排算法············43

2.3.3 5G网络切片使能智能电网···································53

2.3.4  应用于NWDAF中的联邦学习技术····················59

第3章  AI与智能物联网······························63

3.1 5G时代IoT海量数据实时处理·························63

3.2  边缘计算与云边协同··········································65

3.2.1  边缘计算·················65

3.2.2  云边协同·················67

3.3  应用于智能IoT中的AI技术····························72

3.3.1  联邦迁移学习·························································72

3.3.2 RPnet网络与车牌识别··········································74

3.3.3  对抗生成网络与移动目标检测·····························76

3.3.4 Android手机去中心化的分布式机器学习···········78

3.3.5  “AI+移动警务”················································79

第4章  AI与5G网络多量纲计费················80

4.1 5G时代变得日益复杂的网络计费····················80

4.2 5G多量纲计费概述············································82

4.2.1  与4G计费量纲对标··············································83

4.2.2 5G计费因子确定···················································85

4.2.3 5G计费欺诈预防···················································86

4.2.4 5G流量异常监测···················································87

4.3  应用于智能计费中的AI技术····························89

4.3.1 ST-DenNetFus算法与网络需求弹性分析············89

4.3.2  强化学习(RL)与客户意图分析························92

第二篇? 客户与管理篇

第5章  AI与客户体验管理··························98

5.1  客户感知网络质量与客观KPI指标差异··········98

5.2 CEM概述···························································102

5.2.1 CEM基本概念·····················································102

5.2.2  客户网络体验感知量化·······································104

5.2.3 CEMC与端到端客户服务体验改善··················106

5.3  应用于CEM中的AI技术·······························108

5.3.1 ADS算法与用户网络感知原因定位··················109

5.3.2 Chatbot技术与客服体验优化·····························111

5.3.3  基于KDtree、LSTM以及多算法融合的网络容量预测··································113

5.3.4 NPS度量与用户业务感知提升··························114

第6章  AI与客户关系管理(CRM)·········118

6.1 5G需求差异化与服务精准化··························118

6.2 CRM概述··························································120

6.2.1 CRM基本概念·····················································120

6.2.2 AI注智客户差异化服务营销······························121

6.3  应用于CRM中的AI技术·······························122

6.3.1 BERT技术在客服NLP中的应用······················122

6.3.2  基于用户单侧通话记录检测的诈骗电话识别···················································127

6.3.3  应用于用户差异化营销中的人脸识别应用技术···············································131

6.3.4  应用于户外广告屏的人体属性识别技术···········134

6.3.5 MPMD加权回归方法在客户画像中的应用实现··············································139

6.3.6  “CRNN+OpenCV”与用户身份证信息自动录入···········································146

6.3.7  基于OCR识别的用户签名信息核对·················148

6.3.8  基于中心性和图相似性算法的智能推荐应用···················································148

6.3.9  基于LDA和MLLT的语音识别特征变换矩阵估计方法································150

6.3.10  基于MFCC和Kaldi-chain声学模型的语音情绪分析···································153

第7章  AI与流程管理································156

7.1  智能流程管理与企业降本增效························156

7.2 AIRPA助力数字化转型····································157

7.2.1 RPA概述··············157

7.2.2 RPA开发运行流程··············································161

7.2.3 RPA开发工具······················································163

7.2.4 RPA管控调度······················································164

7.2.5 RPA任务执行引擎··············································166

7.3  应用于智能流程管理中的AI技术··················167

7.3.1 YOLO模型检测和分类票据·······························167

7.3.2  用OpenCV去除印章···········································169

7.3.3 CRNN识别票据关键信息···································170

7.3.4  基于模板的OCR识别·········································171

第8章  AI与商业智能································173

8.1 5G与运营商业务决策和业务流程优化··········173

8.2  构建基于通信AI的全面战略管理决策体系··················································176

8.3  应用于智能决策中的AI技术··························177

8.3.1  纳什均衡算法与携号转网最优市场决策···········177

8.3.2  Transfer Learning(迁移学习)技术与客户携转风险识别······························183

8.3.3  基于多源指标关联分析的业务沙盘推演···········186

8.3.4  基于社群发现的用户转网预警分析···················192

第三篇? 运维与安全篇

第9章  AI与网络智能运维························198

9.1 5G网络复杂化与运维模式创新······················198

9.2 AIOps概述·························································200

9.2.1 AIOps概念与关键业务流程·······························200

9.2.2 AIOps与智能运维学件·······································202

9.3  应用于智能运维中的AI技术··························204

9.3.1  基于动态阈值的网络运维异常检测···················204

9.3.2  基于DBSCAN和Apriori算法的传输网告警根因定位···································209

9.3.3  集成学习算法与网络故障预测···························214

9.3.4  时序算法与网络黄金指标预测···························216

9.3.5  基于异构知识关联的运维知识图谱构建···········218

第10章AI与机房智慧管控·······················221

10.1 5G时代的中心机房智慧管控························221

10.2  机房资源调度与监控管理概述······················223

10.2.1  机房环境物理指标·············································223

10.2.2  “IoT+AI”辅助机房管理自动化·····················224

10.2.3  机房安防布控与违规预警·································225

10.3  应用于机房智能化中的AI技术····················225

10.3.1  机器学习方法辅助数据中心降低能源消耗·····················································225

10.3.2 Faster-RCNN目标检测算法监控机柜资源占用··············································229

10.3.3  基于计算机视觉方法的机房火情监测·············233

第11章AI与智能安防······························235

11.1  “5G+AI”安防发展趋势·······························236

11.2  应用于智能安防中的5G技术·······················239

11.2.1  无线视频监控部署·············································239

11.2.2  三域一体立体化防控·········································241

11.2.3  海量数据实时响应·············································242

11.3  应用于智能安防中的AI技术························244

11.3.1 AI安防模型························································244

11.3.2 AI服务实现························································250

11.3.3  资源混编调度·····················································252

第12章5G时代的AI能力平台化············255

12.1 AI平台建设与能力沉积·································255

12.2 AI平台建设理念与思路·································256

12.3  AI平台建设功能设计····································261

12.3.1  云化引擎设计·····················································261

12.3.2 API算法体系······················································262

12.3.3 AI能力生产方式················································262

12.3.4 AI能力输出方式················································265

12.3.5  与生产环境对接·················································266

12.4  AI平台建设的技术设计·································267

参考文献······················································269

作者简介

编辑推荐

来自国内大型电信软件与服务提供商的实战分享;

揭秘5G时代如何将AI技术赋能生产与运营管理;

涉及5G智能网络、5G切片、CEM、CRM、智能安防等典型业务场景;

分享AI中台建设的理念和设计建议

作者寄语

亚信科技(中国)有限公司(简称亚信科技,股票代码01675.HK)创建于1993年,是国内领先的软件产品、解决方案和服务提供商,致力于成为5G时代大型企业数字化转型的使能者。

公司积极拥抱5G、云计算、大数据、AI、物联网等先进的技术,依据“一巩固、三发展”的战略决策,依托产品、服务、运营和集成的能力,在传统业务方面,以5G为契机,全面布局,提升效能,巩固BSS市场的领导地位;在新兴业务方面,5G OSS网络智能化、DSaaS数字化运营服务、企业上云及垂直行业领域快速规模化发展。同时,公司将与业界伙伴共同建设生态体系,持续推动商业模式转型,为企业数字化转型和产业可持续发展贡献力量。

亚信科技拥有行业领先的研发能力和丰富的电信级软件产品,包括客户关系管理、计费账 务、大数据、物联网及5G网络智能化产品。大型企业客户来自金融、交通、邮政、能源、广电、零售、政务等行业。

电子资料

www.luweidong.cn

下一个