-
Notifications
You must be signed in to change notification settings - Fork 0
/
note.out
68 lines (68 loc) · 5.15 KB
/
note.out
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
\BOOKMARK [0][-]{chapter.1}{Mathematical Foundation}{}% 1
\BOOKMARK [1][-]{section.1.1}{Advanced math}{chapter.1}% 2
\BOOKMARK [2][-]{subsection.1.1.1}{Taylor formula}{section.1.1}% 3
\BOOKMARK [1][-]{section.1.2}{Probability theory and mathematical statistics}{chapter.1}% 4
\BOOKMARK [2][-]{subsection.1.2.1}{How to get expected value and variance?}{section.1.2}% 5
\BOOKMARK [2][-]{subsection.1.2.2}{Discrete probability distribution}{section.1.2}% 6
\BOOKMARK [2][-]{subsection.1.2.3}{Continuous probability distribution}{section.1.2}% 7
\BOOKMARK [2][-]{subsection.1.2.4}{Sample mean and sample variance}{section.1.2}% 8
\BOOKMARK [0][-]{chapter.2}{Machine Learning}{}% 9
\BOOKMARK [1][-]{section.2.1}{Linear regression}{chapter.2}% 10
\BOOKMARK [2][-]{subsection.2.1.1}{Unitary linear regression}{section.2.1}% 11
\BOOKMARK [1][-]{section.2.2}{Linear Gression Least mean square \(LMS\)}{chapter.2}% 12
\BOOKMARK [1][-]{section.2.3}{Logistic regression \(LR\)}{chapter.2}% 13
\BOOKMARK [2][-]{subsection.2.3.1}{Form}{section.2.3}% 14
\BOOKMARK [1][-]{section.2.4}{Naive Bayes}{chapter.2}% 15
\BOOKMARK [2][-]{subsection.2.4.1}{Prior and posterior}{section.2.4}% 16
\BOOKMARK [2][-]{subsection.2.4.2}{Naive Bayesian Classifier \(NBC\)}{section.2.4}% 17
\BOOKMARK [2][-]{subsection.2.4.3}{Parameter estimation of NBC by maximum likelihood estimation \(MLE\)}{section.2.4}% 18
\BOOKMARK [1][-]{section.2.5}{Regularization}{chapter.2}% 19
\BOOKMARK [2][-]{subsection.2.5.1}{L0}{section.2.5}% 20
\BOOKMARK [2][-]{subsection.2.5.2}{L1 \(lasso regularization\)}{section.2.5}% 21
\BOOKMARK [2][-]{subsection.2.5.3}{L2 \(ridge regression or weight decay\)}{section.2.5}% 22
\BOOKMARK [1][-]{section.2.6}{Perceptron}{chapter.2}% 23
\BOOKMARK [2][-]{subsection.2.6.1}{Why can't perceptron solve XOR problem?}{section.2.6}% 24
\BOOKMARK [2][-]{subsection.2.6.2}{Definition of perceptron}{section.2.6}% 25
\BOOKMARK [2][-]{subsection.2.6.3}{Learning Algorithm}{section.2.6}% 26
\BOOKMARK [2][-]{subsection.2.6.4}{Dual form of perceptron learning algorithm}{section.2.6}% 27
\BOOKMARK [1][-]{section.2.7}{Support vector machine \(SVM\)}{chapter.2}% 28
\BOOKMARK [2][-]{subsection.2.7.1}{Form}{section.2.7}% 29
\BOOKMARK [2][-]{subsection.2.7.2}{Lagrange duality}{section.2.7}% 30
\BOOKMARK [2][-]{subsection.2.7.3}{Solution of SVM}{section.2.7}% 31
\BOOKMARK [2][-]{subsection.2.7.4}{Kernel}{section.2.7}% 32
\BOOKMARK [1][-]{section.2.8}{How to get the update rule of parameters of backpropagation in gradient descent algorithm?}{chapter.2}% 33
\BOOKMARK [1][-]{section.2.9}{Gaussian mixture model \(GMM\)}{chapter.2}% 34
\BOOKMARK [2][-]{subsection.2.9.1}{Maximum likelihood estimation}{section.2.9}% 35
\BOOKMARK [2][-]{subsection.2.9.2}{GMM and EM}{section.2.9}% 36
\BOOKMARK [1][-]{section.2.10}{Principal components analysis \(PCA\)}{chapter.2}% 37
\BOOKMARK [2][-]{subsection.2.10.1}{Maximum variance theory}{section.2.10}% 38
\BOOKMARK [2][-]{subsection.2.10.2}{PCA}{section.2.10}% 39
\BOOKMARK [1][-]{section.2.11}{Hidden Markov model \(HMM\)}{chapter.2}% 40
\BOOKMARK [2][-]{subsection.2.11.1}{Definition}{section.2.11}% 41
\BOOKMARK [2][-]{subsection.2.11.2}{Three basic problems}{section.2.11}% 42
\BOOKMARK [1][-]{section.2.12}{Conditional random field \(CRF\)}{chapter.2}% 43
\BOOKMARK [2][-]{subsection.2.12.1}{Probabilistic undirected graphical model, or Markov random field}{section.2.12}% 44
\BOOKMARK [2][-]{subsection.2.12.2}{Conditional random field}{section.2.12}% 45
\BOOKMARK [1][-]{section.2.13}{Generative model and discriminative model}{chapter.2}% 46
\BOOKMARK [0][-]{chapter.3}{Deep Network}{}% 47
\BOOKMARK [1][-]{section.3.1}{How does backpropagation work?}{chapter.3}% 48
\BOOKMARK [1][-]{section.3.2}{Backpropagation of Convolutional Neural Network}{chapter.3}% 49
\BOOKMARK [2][-]{subsection.3.2.1}{Backpropagation of Fully Connectional Layer}{section.3.2}% 50
\BOOKMARK [2][-]{subsection.3.2.2}{Backpropagation of Convolutional Layer}{section.3.2}% 51
\BOOKMARK [1][-]{section.3.3}{Why does batch normalization work?}{chapter.3}% 52
\BOOKMARK [2][-]{subsection.3.3.1}{Dataset shift and covariate shift}{section.3.3}% 53
\BOOKMARK [2][-]{subsection.3.3.2}{Internal covariate shift}{section.3.3}% 54
\BOOKMARK [2][-]{subsection.3.3.3}{Batch normalization}{section.3.3}% 55
\BOOKMARK [1][-]{section.3.4}{Why does residual learning work?}{chapter.3}% 56
\BOOKMARK [1][-]{section.3.5}{Understanding Deconvolution \046 bilinear interpolation}{chapter.3}% 57
\BOOKMARK [2][-]{subsection.3.5.1}{Why is it called Transposed Convolution?}{section.3.5}% 58
\BOOKMARK [2][-]{subsection.3.5.2}{Bilinear Interpolation}{section.3.5}% 59
\BOOKMARK [1][-]{section.3.6}{Activation Function}{chapter.3}% 60
\BOOKMARK [2][-]{subsection.3.6.1}{Sigmoid}{section.3.6}% 61
\BOOKMARK [2][-]{subsection.3.6.2}{ReLU}{section.3.6}% 62
\BOOKMARK [1][-]{section.3.7}{Parameter Initialization}{chapter.3}% 63
\BOOKMARK [2][-]{subsection.3.7.1}{Why not initialized with all zeros?}{section.3.7}% 64
\BOOKMARK [2][-]{subsection.3.7.2}{Why not initialized with Standard Gaussian?}{section.3.7}% 65
\BOOKMARK [2][-]{subsection.3.7.3}{xavier}{section.3.7}% 66
\BOOKMARK [2][-]{subsection.3.7.4}{msra}{section.3.7}% 67
\BOOKMARK [0][-]{chapter.4}{Contents specifically referenced}{}% 68