教育>本科研究生>计算机类
边云智能数据分析与应用  

边云智能数据分析与应用  "

作者:沈钧戈等
ISBN:9787121460425
定价:¥55.0
字数:216千字
页数:216
出版时间:2023-08
开本:16开
版次:01-01
装帧:
出版社:电子工业出版社
简介

随着“十四五”规划纲要中提出“协同发展云服务与边缘计算服务”的观点,边云智能已成为未来发展的重要趋势。本书依托于政策大背景,旨在向读者介绍边云智能的基础知识和应用。书中分为四个篇章,第一篇章介绍了边云架构的骨架和基础概念,第二篇章介绍了人工智能算法和深度学习模型,第三篇章介绍了云端决策算法和边缘端轻量化算法,第四篇章介绍了边云智能在智慧教育领域的应用。本书可以使读者了解边云计算的基本概念和原理逻辑,熟悉基本的人工智能计算方法和数据分析的逻辑和运用场景。通过数据科学的思路和方法,读者可以将无人系统的数据智能化应用提升,并培养数据导向思维方式,为未来学习智能无人系统科学与技术学科打下基础。 本书目标明确,技术先进,强调课程思政和润物无声的教育理念,旨在提高学生的数据科学素养和“用数据”的能力。本书面向智能无人系统科学与技术专业的研究生,涵盖人工智能、大数据分析、数据挖掘和边云计算等学科,具有交叉性的特点。同时,资深从业者也可将其作为参考书籍。

前言

2019 年开始,边缘计算备受关注,一度引起了资本市场的投资热潮,因此很多人把 2019 年称为边缘计算的元年。工业互联网的大力推动、5G 大规模商用的持续酝酿等因素,让整个产业对 IT(Information Technology)和 OT(Operation Technology)的深度融合充满信心和期待,在这种情况下,边缘计算受到越来越多人的关注。 为了褒扬边缘计算,有人将边缘计算称为云计算的“终结者”,并刻意将云计算放在边缘计算的对立面,然而从技术或商业演进的现实情景来看,边缘计算其实更多体现为云计算向终端和用户端延伸时产生的新的解决方案,并且经过时间的验证,云 计算和边缘计算的关系更加清晰。简单来说,云计算的计算效率更高,而边缘计算可以提供低时延,提升用户体验,这种情况下只有两者互为补充,才能更好地处理现实生活中的各种复杂的需求场景,最大化云计算和边缘计算的应用价值,实现边缘端与云端的智能协同。 同时受益于算法、算力和数据集等方面的发展,人工智能技术得到了突飞猛进的发展,在安防、交通、工业、农业等各行业得到广泛应用。随着 5G、物联网时代的到来,为借助边缘侧数据采集便利、实时处理计算等特点,人工智能技术逐步从中心云向边缘下沉,通过将模型在边缘和云端进行协同推理和训练,解决人工智能落地“最后一公里”问题,边缘智能应运而生。 以自动驾驶为例,其在未来的计算模式便是边缘计算与云计算相结合,部署于边缘端的自动驾驶专用芯片会实时感知传感器数据并立刻处理、做决策;同时,这些处理之后的数据会被汇聚在云端,进行大数据分析、模型搭建和编辑、大规模仿真模拟,进行深度分析和机器学习,并对已有的边缘端设备实现更新和升级,使边缘端设备更智能。因此,“算法+芯片+云计算”的模式构成了未来自动驾驶的三大核心支点。 再以物联网为例,阿里云发布的边缘计算产品 Link Edge 通过赋予家庭网关计算能边云智能数据分析与应用 ,即便是在断网的情况下,例如生物识别门锁、机器人等都能正常运作。但是,如果加上云计算,基于云端的大数据分析和判断,在联动的前提下,整个家庭场景的智能设备将变得更为个性化和智能化,例如关上门的时候,扫地机器人就开始运作等。 由此可以看出边缘端设备主要应用在前端,负责数据的实时采集、计算和处理,而云端提供大数据处理、大数据存储、应用程序开发、机器学习和人工智能等方面的强大处理能力,同时,在云端的应用设计、开发、测试、部署、管理等功能也是满足各种应用需求的关键,因此如何合理发挥边缘端与云端各自的优势、实现边云智能协同是目前亟待解决的问题。 1.本书特色 本书面向智能无人系统科学与技术学科的研究生,具有人工智能、大数据分析、数据挖掘、边云计算学科交叉的特点,旨在扩大学生视野,提高科学素养,培养多元化思维方式,拓宽知识面,促进学科交叉,读者通过学习本书: (1)能够了解边云计算的基本概念和原理逻辑; (2)了解基本的人工智能计算方法,熟悉数据分析的基本逻辑和运用场景; (3)能够运用一定的工具采用数据科学的思路和方法对无人系统进行数据方面的智能化应用提升,来研究和分析无人系统学科的数据驱动问题; (4)培养数据导向思维方式,为将来的智能无人系统科学与技术学科的进一步学习打下基础。 本书目标明确、技术先进。在编写上,本书以提高学生的数据科学素养与“用数据”的能力为出发点;在内容的选取上,本书充分满足从事无人系统相关工作应具有的数据分析建模知识及应用技能需求。 2.内容简介 边云智能协同技术的发展目前已势不可挡,“十四五”纲要文件正式提出了“协同发展云服务与边缘计算服务”的观点,这意味着边云智能协同技术成为了未来发展的重要趋势。 目前市面上系统讲述边云智能协同及其关键技术的书籍较少,因此本书旨在向读者系统且全面地介绍边云智能体系的理论知识和实战案例分析。为了系统地为读者讲明边云智能体系及应用,本书将以 4 个篇章逐步向读者阐述相关内容,分别为:①基础架构篇;②智能理论篇;③核心技术篇;④应用实践篇。接下来对各篇章进行简要叙述。 (1)基础架构篇介绍边云智能体系的骨架——边云架构。基础架构篇首先阐述了边云智能产生的大背景、基础概念、发展过程及应用场景,让读者对边云智能体系有初步的认识,接着向读者介绍边云架构系统的设计理念、度量方式与应用,让读者对边缘端与云端的智能协同模式有更深刻的认识。 (2)智能理论篇介绍边云智能体系的身体——人工智能算法。人工智能算法是赋能边云智能体系之所以“智能”的关键内容,而深度学习算法在如今人工智能领域占举足轻重的地位,因此,该篇首先介绍了人工智能与深度学习相关概念及经典的神经网络模型,接着介绍深度学习的两个重要应用领域:自然语言处理与计算机视觉,其分别代表着边云智能体系的“嘴巴”与“眼睛”。 (3)核心技术篇介绍边云智能体系的“手足”与“大脑”——云端决策与边缘端轻量化。核心技术篇首先介绍边云智能体系中云端“智能”的关键,即云端决策算法,接着对边缘端智能落地的关键算法进行了介绍,即边缘端轻量化算法。 (4)应用实践篇介绍边云智能体系的应用实践,它以以上三个篇章为依托,向读者详细讲述边云智能在智慧教育领域的作用,包括边云智能体系框架的设计理念、边缘端智能与云智能的算法原理。 3.使用建议 如果读者之前没有边云架构或机器学习的学习经验,我们强烈建议读者先从本书的基础架构篇或智能理论篇开始学习,然后再在此基础上开始核心技术篇的学习,最后从应用实践篇中体会边云智能的应用。

目录

第 1 章 绪论 ····························································································1 1.1 边云智能产生的大背景····································································1 1.1.1 新一代信息技术的快速发展·····················································2 1.1.2 国家政策的支持与引导···························································6 1.2 边云智能······················································································7 1.3 边云智能的发展·············································································9 1.3.1 边云智能的三大发展阶段························································9 1.3.2 城市大脑··········································································.11 1.4 “智能+”新潮头··········································································.13 1.4.1 “智能+”技术新融合···························································.13 1.4.2 多维度场景应用·································································.14 本章习题··························································································.15 第 2 章 边云架构 ···················································································.16 2.1 系统工程方法论··········································································.17 2.1.1 概述 ················································································.17 2.1.2 基本方法··········································································.17 2.2 边云智能体系架构模型·································································.20 2.2.1 概念框架··········································································.20 2.2.2 层次结构··········································································.22 2.3 协同模式···················································································.23 2.3.1 “云-边”协同 ····································································.24 2.3.2 “边-边”协同 ····································································.25 2.3.3 “边-端”协同 ····································································.27 2.3.4 “云-边-端”协同 ································································.28 2.3.5 度量指标··········································································.28 2.4 边云智能架构应用·······································································.30 2.4.1 “云-边-端”区块链 ·····························································.30 2.4.2 “云-边-端”一体化机器人系统 ··············································.32 本章习题··························································································.33 第 3 章 深度学习 ···················································································.35 3.1 深度学习概念·············································································.36 3.1.1 人工智能与机器学习···························································.36 3.1.2 深度学习··········································································.37 3.1.3 神经网络··········································································.39 3.2 前馈神经网络·············································································.39 3.2.1 感知机模型·······································································.39 3.2.2 反向传播··········································································.42 3.2.3 卷积神经网络····································································.44 3.2.4 几种典型的卷积神经网络·····················································.47 3.3 反馈神经网络·············································································.50 3.3.1 循环神经网络····································································.50 3.3.2 长短期神经网络·································································.53 3.4 Transformer 神经网络 ···································································.56 3.4.1 编码器单元与解码器单元·····················································.58 3.4.2 多头注意力机制·································································.59 3.4.3 非参位置编码····································································.60 本章习题··························································································.61 第 4 章 自然语言处理 ·············································································.62 4.1 自然语言处理概述·······································································.63 4.1.1 自然语言处理简介······························································.63 4.1.2 自然语言处理的发展历史·····················································.74 4.1.3 自然语言处理的应用及面临的挑战·········································.76 4.2 文本挖掘···················································································.79 4.2.1 文本挖掘简介····································································.79 4.2.2 文本挖掘算法····································································.81 4.3 机器翻译···················································································.87 4.3.1 机器翻译简介····································································.87 4.3.2 机器翻译算法····································································.89 4.4 自动问答系统·············································································.93 4.4.1 自动问答系统简介······························································.93 4.4.2 自动问答系统模型······························································.95 4.5 语音识别···················································································101 4.5.1 语音识别简介····································································102 4.5.2 语音识别算法····································································103 本章习题··························································································105 第 5 章 计算机视觉 ················································································107 5.1 计算机视觉概述··········································································107 5.1.1 计算机视觉简介·································································108 5.1.2 计算机视觉的发展历史························································109 5.1.3 计算机视觉的应用及面临的挑战···········································.110 5.2 图像分类··················································································.114 5.2.1 图像分类简介···································································.114 5.2.2 图像分类算法···································································.115 5.3 目标检测··················································································.119 5.3.1 目标检测简介···································································.119 5.3.2 目标检测算法····································································120 5.4 图像分割···················································································123 5.4.1 图像分割简介····································································123 5.4.2 图像分割算法····································································124 5.5 目标跟踪···················································································125 5.5.1 目标跟踪简介····································································126 5.5.2 目标跟踪算法····································································126 本章习题··························································································128 第 6 章 边缘轻量化 ················································································129 6.1 边缘轻量化的简介·······································································129 6.1.1 边缘端对轻量化的需求························································129 6.1.2 什么是边缘轻量化······························································130 6.2 模型压缩方法·············································································131 6.2.1 量化和二值化····································································131 6.2.2 网络剪枝··········································································131 6.2.3 低秩因子分解····································································132 6.2.4 参数共享··········································································133 6.2.5 蒸馏学习··········································································133 6.2.6 加速网络设计····································································134 6.3 模型压缩举例·············································································137 6.3.1 知识蒸馏··········································································137 6.3.2 深度压缩··········································································139 6.3.3 MNASNet ·········································································143 本章习题··························································································145 第 7 章 云端决策 ···················································································146 7.1 云端决策简介·············································································147 7.1.1 云端决策的重要性······························································147 7.1.2 云端决策的特点·································································147 7.2 云端决策——大数据挖掘······························································149 7.2.1 回归分析··········································································149 7.2.2 聚类 ················································································150 7.2.3 关联规则··········································································152 7.3 云端决策——推荐算法·································································154 7.3.1 基于统计的推荐算法···························································155 7.3.2 基于协同过滤的推荐系统·····················································155 7.3.3 基于内容的推荐系统···························································156 7.3.4 基于关联规则的推荐系统·····················································158 7.3.5 基于网络结构的推荐系统·····················································158 本章习题··························································································159 第 8 章 边云智能赋能智慧教室 ·································································160 8.1 智慧教室的形成背景与边云框架·····················································161 8.1.1 智慧教室政策支持与特征分析 ··············································162 8.1.2 基于边云智能的智慧教室框架 ··············································164 8.1.3 基于边云智能建设的智慧教室目标愿景 ··································166 8.2 智慧教室的边缘端感知技术与应用··················································166 8.2.1 无感考勤、表情感知与异常行为识别 ·····································167 8.2.2 边缘端感知模型的压缩与轻量化 ···········································172 8.3 智慧教室的云端决策技术与应用·····················································174 8.3.1 “教育大脑”大数据分析决策方法··········································174 8.3.2 个性化推荐、学习评价与师生互动应用 ··································176 8.4 “边云智能+”前景展望·································································178 8.4.1 边云智能赋能智慧交通························································178 8.4.2 边云智能赋能智慧安防························································185 本章习题··························································································190 习题答案································································································191

作者简介

编辑推荐

作者寄语

电子资料

www.luweidong.cn

下一个