第1 章 物聯(lián)網(wǎng)技術(shù)基礎(chǔ)···········································································.001
1.1 物聯(lián)網(wǎng)發(fā)展現(xiàn)狀··········································································.001
1.1.1 美國·················································································.001
1.1.2 歐盟·················································································.002
1.1.3 日本·················································································.002
1.1.4 韓國·················································································.003
1.1.5 中國·················································································.003
1.2 物聯(lián)網(wǎng)體系架構(gòu)··········································································.004
1.2.1 感知層·············································································.005
1.2.2 網(wǎng)絡(luò)層·············································································.005
1.2.3 應(yīng)用層·············································································.006
1.3 物聯(lián)網(wǎng)信息的三大特性·······························································.007
1.3.1 高敏感性··········································································.007
1.3.2 高可靠性··········································································.007
1.3.3 高關(guān)聯(lián)性··········································································.007
1.4 物聯(lián)網(wǎng)體系架構(gòu)的發(fā)展·······························································.008
1.5 小結(jié)·······························································································011
第2 章 物聯(lián)網(wǎng)安全架構(gòu)··········································································.012
2.1 引言·····························································································.012
2.2 物聯(lián)網(wǎng)典型安全事件···································································.012
2.2.1 事件回顧··········································································.012
2.2.2 事件分析··········································································.016
2.3 物聯(lián)網(wǎng)安全問題分類···································································.017
2.3.1 互聯(lián)網(wǎng)引入的安全問題···················································.017
2.3.2 物聯(lián)網(wǎng)場景下的互聯(lián)網(wǎng)“安全”問題····························.017
2.3.3 物聯(lián)網(wǎng)引入的安全問題···················································.017
2.3.4 物聯(lián)網(wǎng)自身的安全問題···················································.018
2.4 物聯(lián)網(wǎng)安全威脅分析···································································.018
2.4.1 感知層安全威脅······························································.019
2.4.2 網(wǎng)絡(luò)層安全威脅······························································.020
2.4.3 應(yīng)用層安全威脅······························································.022
2.5 物聯(lián)網(wǎng)安全需求分析···································································.023
2.5.1 縱橫聯(lián)動的一體化安全保障支撐····································.024
2.5.2 感知層傳感器設(shè)備的身份鑒別與數(shù)據(jù)防護(hù)····················.026
2.5.3 網(wǎng)絡(luò)層異構(gòu)網(wǎng)絡(luò)規(guī)模化安全互聯(lián)與全網(wǎng)統(tǒng)一監(jiān)管·········.027
2.5.4 應(yīng)用層數(shù)據(jù)多域安全共享···············································.028
2.6 物聯(lián)網(wǎng)安全體系框架···································································.029
2.6.1 技術(shù)體系··········································································.029
2.6.2 物聯(lián)網(wǎng)系統(tǒng)的安全信息流···············································.036
2.7 小結(jié)·····························································································.037
第3 章 物聯(lián)網(wǎng)安全認(rèn)證技術(shù)···································································.039
3.1 引言·····························································································.039
3.2 身份認(rèn)證方式··············································································.040
3.2.1 基于秘密信息的認(rèn)證方式···············································.040
3.2.2 基于信物的認(rèn)證方式·······················································.041
3.2.3 基于密鑰的認(rèn)證方式·······················································.041
3.2.4 基于生物特征的認(rèn)證方式···············································.042
3.3 物聯(lián)網(wǎng)身份認(rèn)證的特點(diǎn)·······························································.043
3.3.1 輕量級·············································································.043
3.3.2 非對稱·············································································.043
3.3.3 復(fù)雜性·············································································.043
3.4 幾種物聯(lián)網(wǎng)身份認(rèn)證方式···························································.044
3.4.1 基于RFID 的物聯(lián)網(wǎng)身份認(rèn)證方式·································.044
3.4.2 基于傳感網(wǎng)絡(luò)的物聯(lián)網(wǎng)認(rèn)證方式····································.044
3.4.3 基于藍(lán)牙的感知網(wǎng)絡(luò)認(rèn)證方式·······································.045
3.4.4 基于生物特征識別的認(rèn)證方式·······································.045
3.5 基于生物特征的物聯(lián)網(wǎng)身份認(rèn)證方法········································.047
3.5.1 生物特征身份認(rèn)證流程···················································.047
3.5.2 指紋識別··········································································.049
3.5.3 人臉識別··········································································.049
3.5.4 虹膜識別··········································································.051
3.5.5 指靜脈識別······································································.051
3.5.6 聲紋識別··········································································.052
3.6 基于深度學(xué)習(xí)的聲紋識別技術(shù)····················································.053
3.6.1 概述·················································································.053
3.6.2 聲紋識別的工作原理·······················································.054
3.6.3 聲紋識別的流程······························································.055
3.6.4 聲紋識別技術(shù)的三次突破···············································.059
3.6.5 基于深度學(xué)習(xí)的典型聲紋識別算法································.060
3.6.6 聲紋識別應(yīng)用趨勢··························································.063
3.7 小結(jié)·····························································································.066
本章參考文獻(xiàn)······················································································.066
第4 章 卷積神經(jīng)網(wǎng)絡(luò)技術(shù)······································································.068
4.1 卷積運(yùn)算······················································································.069
4.2 動機(jī)·····························································································.072
4.3 池化·····························································································.077
4.4 將卷積與池化作為一個無限強(qiáng)的先驗(yàn)········································.082
4.5 基本卷積函數(shù)的變體···································································.083
4.6 結(jié)構(gòu)化輸出··················································································.093
4.7 數(shù)據(jù)類型······················································································.094
4.8 高效的卷積算法··········································································.095
4.9 隨機(jī)或無監(jiān)督的特征···································································.096
4.10 小結(jié)···························································································.097
本章參考文獻(xiàn)······················································································.098
第5 章 序列建模：循環(huán)和遞歸網(wǎng)絡(luò)·······················································.101
5.1 展開計算圖··················································································.102
5.2 RNN·····························································································.105
5.2.1 Teacher Forcing 和輸出循環(huán)網(wǎng)絡(luò)····································.109
5.2.2 計算RNN 的梯度······························································111
5.2.3 作為有向圖模型的循環(huán)網(wǎng)絡(luò)·············································113
5.2.4 基于上下文的RNN 序列建模···········································117
5.3 雙向RNN ······················································································119
5.4 基于編碼-解碼的序列到序列架構(gòu)···············································.121
5.5 深度RNN ····················································································.123
5.6 遞歸神經(jīng)網(wǎng)絡(luò)··············································································.124
5.7 長期依賴的挑戰(zhàn)··········································································.126
5.8 回聲狀態(tài)網(wǎng)絡(luò)··············································································.128
5.9 滲漏單元和其他多時間尺度的策略············································.130
5.9.1 時間維度的跳躍連接·······················································.130
5.9.2 滲漏單元和一系列時間尺度···········································.131
5.9.3 刪除連接··········································································.131
5.10 長短期記憶和其他門控RNN ····················································.132
5.10.1 長短期記憶····································································.133
5.10.2 其他門控RNN·······························································.135
5.11 優(yōu)化長期依賴············································································.136
5.11.1 截斷梯度········································································.136
5.11.2 引導(dǎo)信息流的正則化·····················································.138
5.12 外顯記憶····················································································.139
5.13 小結(jié)···························································································.142
本章參考文獻(xiàn)······················································································.142
第6 章 深度學(xué)習(xí)框架和平臺的分析與對比············································.148
6.1 概述·····························································································.148
6.2 深度學(xué)習(xí)框架··············································································.151
6.2.1 TensorFlow·······································································.152
6.2.2 Caffe ················································································.156
6.2.3 PyTorch ············································································.162
6.2.4 CNTK···············································································.164
6.2.5 MXNet ·············································································.166
6.3 深度學(xué)習(xí)框架的分析與對比·······················································.169
6.3.1 總體分析··········································································.169
6.3.2 深度學(xué)習(xí)框架的對比·······················································.170
6.3.3 深度學(xué)習(xí)框架對硬件的利用情況····································.178
6.4 深度學(xué)習(xí)平臺··············································································.180
6.4.1 華為深度學(xué)習(xí)服務(wù)DLS ··················································.180
6.4.2 阿里深度學(xué)習(xí)開發(fā)平臺X-DeepLearning ························.184
6.4.3 百度深度學(xué)習(xí)開發(fā)平臺PAddle·······································.189
6.4.4 幾種平臺的對比······························································.191
6.5 小結(jié)·····························································································.192