Download: HTTP: // gqylpy / DI / 194
"Layman deep learning: Principles and analyze python practice .pdf" PDF HD full version - Download
"deep learning in simple terms: the principle of analysis and Python Practice" describes the depth of learning-related theory and application of book is divided into three parts, the first part mainly reviews the development history of deep learning, and Theano use; the second part explains the basic knowledge related to the depth of learning, including linear algebra, probability theory, probabilistic graphical models, machine learning and optimization algorithms to; in the third part , a learning model for the depth of a number of core, such as from the encoder, restricted Boltzmann machine recursive convolutional neural networks and neural network analysis and detailed explanation of the principle, and the corresponding specific for different models application.
"Layman deep learning: Principles and analysis of Python practice" for a certain higher mathematics, machine learning and Python-based programming in school, university researchers or engage in deep learning in the enterprise engineers use the principles of the model and difficult book in-depth analysis, in the back of each chapter provides detailed references, the reader may be more in-depth study of the relevant details. Theory and practice, "easy to understand the depth of learning: Principles and Python profiling practices" for popular models are given a corresponding application, readers can also download and view on Github in "deep learning in simple terms: the principle of analysis and Python practice" Code ( https://github.com/innovation-cat/DeepLearningBook ).
About the Author
Huang port, graduated from Tsinghua University in 2012, received a master's degree in school active in other programming competition TopCoder community. Now includes personalized recommendations, natural language processing and large-scale similarity optimization, especially for deep learning in-depth study in the application recommendation system for senior Tencent basic research, research, and applied for more than 10 domestic related items patent.
table of Contents
Part 1 1 Summary
1 Introduction 2
1.1 Relations artificial intelligence, machine learning and deep learning of 3
1.1.1 Artificial Intelligence - Machine Reasoning 4
1.1.2 机器学习——数据驱动的科学 5
1.1.3 深度学习——大脑的仿真 8
1.2 深度学习的发展历程 8
1.3 深度学习技术概述 10
1.3.1 从低层到高层的特征抽象 11
1.3.2 让网络变得更深 13
1.3.3 自动特征提取 14
1.4 深度学习框架 15
2 Theano 基础 19
2.1 符号变量 20
2.2 符号计算的抽象——符号计算图模型 23
2.3 函数 26
2.3.1 函数的定义 26
2.3.2 Logistic回归 27
2.3.3 函数的复制 29
2.4 条件表达式 31
2.5 循环 32
2.6 共享变量 39
2.7 配置 39
2.7.1 通过THEANO_FLAGS配置 40
2.7.2 通过. theanorc文件配置 41
2.8 常用的Debug技巧 42
2.9 小结 43
第2 部分 数学与机器学习基础篇 45
3 线性代数基础 46
3.1 标量、向量、矩阵和张量 46
3.2 矩阵初等变换 47
3.3 线性相关与向量空间 48
3.4 范数 49
3.4.1 向量范数 49
3.4.2 矩阵范数 53
3.5 特殊的矩阵与向量 56
3.6 特征值分解 57
3.7 奇异值分解 58
3.8 迹运算 60
3.9 样例：主成分分析 61
4 概率统计基础 64
4.1 样本空间与随机变量 65
4.2 概率分布与分布函数 65
4.3 一维随机变量 66
4.3.1 离散型随机变量和分布律 66
4.3.2 连续型随机变量和概率密度函数 67
4.4 多维随机变量 68
4.4.1 离散型二维随机变量和联合分布律 69
4.4.2 连续型二维随机变量和联合密度函数 69
4.5 边缘分布 70
4.6 条件分布与链式法则 71
4.6.1 条件概率 71
4.6.2 链式法则 73