《深度學習：原理與套用實踐》是2016年12月電子工業出版社出版的圖書，作者是張重生。
基本介紹
- 中文名：深度學習：原理與套用實踐
- 作者：張重生
- 出版社：電子工業出版社
- 出版時間：2016年12月
- 頁數：232 頁
- 定價：48 元
- 開本：16 開
- ISBN：9787121304132
圖書目錄,作者簡介,內容簡介,
圖書目錄
本書全面、系統地介紹深度學習相關的技術，包括人工神經網路，卷積神經網路，深度學習平台及原始碼分析，深度學習入門與進階，深度學習高級實踐，所有章節均附有源程式，所有實驗讀者均可重現，具有高度的可操作性和實用性。通過學習本書，研究人員、深度學習愛好者，能夠在3 個月內，系統掌握深度學習相關的理論和技術。
作者簡介
張重生，男，博士，教授，碩士生導師，河南大學大數據研究中心、大數據團隊帶頭人。研究領域為大數據分析、深度學習、數據挖掘、資料庫、數據流（實時數據分析）。__eol__博士畢業於 INRIA,France(法國國家信息與自動化研究所)，獲得優秀博士論文榮譽。2010年08月至2011年3月，在美國加州大學洛杉磯分校(UCLA)，計算機系，師從著名的資料庫專家Carlo Zaniolo教授，從事數據挖掘領域的合作研究。 2012-2013,挪威科技大學，ERCIM/Marie-Curie Fellow
內容簡介
目 錄
深度學習基礎篇
第1 章 緒論 ·································································································· 2
1.1 引言 ······································································································· 2
1.1.1 Google 的深度學習成果 ···························································· 2
1.1.2 Microsoft 的深度學習成果························································· 3
1.1.3 國內公司的深度學習成果 ························································· 3
1.2 深度學習技術的發展歷程 ···································································· 4
1.3 深度學習的套用領域 ············································································ 6
1.3.1 圖像識別領域 ············································································· 6
1.3.2 語音識別領域 ············································································· 6
1.3.3 自然語言理解領域 ····································································· 7
1.4 如何開展深度學習的研究和套用開發 ················································· 7
本章參考文獻 ······························································································ 11
第2 章 國內外深度學習技術研發現狀及其產業化趨勢 ······························· 13
2.1 Google 在深度學習領域的研發現狀 ·················································· 13
2.1.1 深度學習在Google 的套用 ······················································ 13
2.1.2 Google 的TensorFlow 深度學習平台 ······································ 14
2.1.3 Google 的深度學習晶片TPU ·················································· 15
2.2 Facebook 在深度學習領域的研發現狀 ·············································· 15
2.2.1 Torchnet ···················································································· 15
2.2.2 DeepText ··················································································· 16
2.3 百度在深度學習領域的研發現狀 ······················································· 17
2.3.1 光學字元識別 ··········································································· 17
2.3.2 商品圖像搜尋 ··········································································· 17
2.3.3 線上廣告 ·················································································· 18
2.3.4 以圖搜圖 ·················································································· 18
2.3.5 語音識別 ·················································································· 18
2.3.6 百度開源深度學習平台MXNet 及其改進的深度語音識別系統Warp-CTC ····· 19
2.4 阿里巴巴在深度學習領域的研發現狀 ··············································· 19
2.4.1 拍立淘 ······················································································ 19
2.4.2 阿里小蜜——智慧型客服Messenger ········································· 20
2.5 京東在深度學習領域的研發現狀 ······················································· 20
2.6 騰訊在深度學習領域的研發現狀 ······················································· 21
2.7 科創型公司（基於深度學習的人臉識別系統） ······························· 22
2.8 深度學習的硬體支撐——NVIDIA GPU ············································ 23
本章參考文獻 ······························································································ 24
深度學習理論篇
第3 章 神經網路 ························································································· 30
3.1 神經元的概念 ······················································································ 30
3.2 神經網路 ····························································································· 31
3.2.1 後向傳播算法 ··········································································· 32
3.2.2 後向傳播算法推導 ··································································· 33
3.3 神經網路算法示例 ·············································································· 36
本章參考文獻 ······························································································ 38
第4 章 卷積神經網路 ················································································· 39
4.1 卷積神經網路特性 ················································································ 39
4.1.1 局部連線 ·················································································· 40
4.1.2 權值共享 ·················································································· 41
4.1.3 空間相關下採樣 ······································································· 42
4.2 卷積神經網路操作 ·············································································· 42
4.2.1 卷積操作 ·················································································· 42
4.2.2 下採樣操作 ·············································································· 44
4.3 卷積神經網路示例：LeNet-5 ····························································· 45
本章參考文獻 ······························································································ 48
深度學習工具篇
第5 章 深度學習工具Caffe ········································································ 50
5.1 Caffe 的安裝 ························································································ 50
5.1.1 安裝依賴包 ·············································································· 51
5.1.2 CUDA 安裝 ·············································································· 51
5.1.3 MATLAB 和Python 安裝 ························································ 54
5.1.4 OpenCV 安裝（可選） ···························································· 59
5.1.5 Intel MKL 或者BLAS 安裝 ····················································· 59
5.1.6 Caffe 編譯和測試 ····································································· 59
5.1.7 Caffe 安裝問題分析 ································································· 62
5.2 Caffe 框架與原始碼解析 ···································································· 63
5.2.1 數據層解析 ·············································································· 63
5.2.2 網路層解析 ·············································································· 74
5.2.3 網路結構解析 ··········································································· 92
5.2.4 網路求解解析 ········································································· 104
本章參考文獻 ···························································································· 109
第6 章 深度學習工具Pylearn2 ································································ 110
6.1 Pylearn2 的安裝 ·················································································· 110
6.1.1 相關依賴安裝 ·········································································· 110
6.1.2 安裝Pylearn2 ·········································································· 112
6.2 Pylearn2 的使用 ·················································································· 112
本章參考文獻 ····························································································· 116
深度學習實踐篇（入門與進階）
第7 章 基於深度學習的手寫數字識別 ······················································ 118
7.1 數據介紹 ···························································································· 118
7.1.1 MNIST 數據集 ········································································ 118
7.1.2 提取MNIST 數據集圖片 ······················································· 120
7.2 手寫字型識別流程 ············································································ 121
7.2.1 模型介紹 ················································································ 121
7.2.2 操作流程 ················································································ 126
7.3 實驗結果分析 ···················································································· 127
本章參考文獻 ···························································································· 128
第8 章 基於深度學習的圖像識別 ····························································· 129
8.1 數據來源 ··························································································· 129
8.1.1 Cifar10 數據集介紹 ································································ 129
8.1.2 Cifar10 數據集格式 ································································ 129
8.2 Cifar10 識別流程 ··············································································· 130
8.2.1 模型介紹 ················································································ 130
8.2.2 操作流程 ················································································ 136
8.3 實驗結果分析 ······················································································ 139
本章參考文獻 ···························································································· 140
第9 章 基於深度學習的物體圖像識別 ······················································ 141
9.1 數據來源 ··························································································· 141
9.1.1 Caltech101 數據集 ·································································· 141
9.1.2 Caltech101 數據集處理 ·························································· 142
9.2 物體圖像識別流程 ············································································ 143
9.2.1 模型介紹 ················································································ 143
9.2.2 操作流程 ················································································ 144
9.3 實驗結果分析 ···················································································· 150
本章參考文獻 ···························································································· 151
第10 章 基於深度學習的人臉識別 ··························································· 152
10.1 數據來源 ························································································· 152
10.1.1 AT&T Facedatabase 資料庫 ·················································· 152
10.1.2 資料庫處理 ··········································································· 152
10.2 人臉識別流程 ·················································································· 154
10.2.1 模型介紹 ·············································································· 154
10.2.2 操作流程 ·············································································· 155
10.3 實驗結果分析 ·················································································· 159
本章參考文獻 ···························································································· 160
深度學習實踐篇（高級套用）
第11 章 基於深度學習的人臉識別——DeepID 算法 ································ 162
11.1 問題定義與數據來源 ······································································ 162
11.2 算法原理 ·························································································· 163
11.2.1 數據預處理 ··········································································· 163
11.2.2 模型訓練策略 ······································································· 164
11.2.3 算法驗證和結果評估 ··························································· 164
11.3 人臉識別步驟 ·················································································· 165
11.3.1 數據預處理 ··········································································· 165
11.3.2 深度網路結構模型 ······························································· 168
11.3.3 提取深度特徵與人臉驗證 ··················································· 171
11.4 實驗結果分析 ·················································································· 174
11.4.1 實驗數據 ··············································································· 174
11.4.2 實驗結果分析 ······································································· 175
本章參考文獻 ···························································································· 176
第12 章 基於深度學習的表情識別 ··························································· 177
12.1 表情數據 ························································································· 177
12.1.1 Cohn-Kanade（CK+）資料庫 ············································· 177
12.1.2 JAFFE 資料庫 ······································································ 178
12.2 算法原理 ························································································· 179
12.3 表情識別步驟 ·················································································· 180
12.3.1 數據預處理 ··········································································· 180
12.3.2 深度神經網路結構模型 ······················································· 181
12.3.3 提取深度特徵及分類 ··························································· 182
12.4 實驗結果分析 ·················································································· 184
12.4.1 實現細節 ·············································································· 184
12.4.2 實驗結果對比 ······································································· 185
本章參考文獻 ···························································································· 188
第13 章 基於深度學習的年齡估計 ··························································· 190
13.1 問題定義 ························································································· 190
13.2 年齡估計算法 ·················································································· 190
13.2.1 數據預處理 ··········································································· 190
13.2.2 提取深度特徵 ······································································· 192
13.2.3 提取LBP 特徵 ····································································· 196
13.2.4 訓練回歸模型 ······································································· 196
13.3 實驗結果分析 ·················································································· 199
本章參考文獻 ···························································································· 199
第14 章 基於深度學習的人臉關鍵點檢測 ················································ 200
14.1 問題定義和數據來源 ······································································ 200
14.2 基於深度學習的人臉關鍵點檢測的步驟 ······································· 201
14.2.1 數據預處理 ··········································································· 201
14.2.2 訓練深度學習網路模型 ······················································· 206
14.2.3 預測和處理關鍵點坐標 ······················································· 207
本章參考文獻 ···························································································· 212
深度學習總結與展望篇
第15 章 總結與展望 ················································································· 214
15.1 深度學習領域當前的主流技術及其套用領域 ······························· 214
15.1.1 圖像識別 ·············································································· 214
15.1.2 語音識別與自然語言理解 ··················································· 215
15.2 深度學習的缺陷 ·············································································· 215
15.2.1 深度學習在硬體方面的門檻較高 ········································ 215
15.2.2 深度學習在軟體安裝與配置方面的門檻較高 ···················· 216
15.2.3 深度學習最重要的問題在於需要海量的有標註的數據作為支撐 ··· 216
15.2.4 深度學習的最後階段竟然變成枯燥、機械、及其耗時的調參工作 ··· 217
15.2.5 深度學習不適用於數據量較小的數據 ································ 218
15.2.6 深度學習目前主要用於圖像、聲音的識別和自然語言的理解 ····· 218
15.2.7 研究人員從事深度學習研究的困境 ···································· 219
15.3 展望 ································································································· 220
本章參考文獻 ···························································································· 220
