-
Notifications
You must be signed in to change notification settings - Fork 0
/
ml.aux
267 lines (267 loc) · 29.1 KB
/
ml.aux
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
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
\relax
\providecommand\hyper@newdestlabel[2]{}
\providecommand\HyperFirstAtBeginDocument{\AtBeginDocument}
\HyperFirstAtBeginDocument{\ifx\hyper@anchor\@undefined
\global\let\oldcontentsline\contentsline
\gdef\contentsline#1#2#3#4{\oldcontentsline{#1}{#2}{#3}}
\global\let\oldnewlabel\newlabel
\gdef\newlabel#1#2{\newlabelxx{#1}#2}
\gdef\newlabelxx#1#2#3#4#5#6{\oldnewlabel{#1}{{#2}{#3}}}
\AtEndDocument{\ifx\hyper@anchor\@undefined
\let\contentsline\oldcontentsline
\let\newlabel\oldnewlabel
\fi}
\fi}
\global\let\hyper@last\relax
\gdef\HyperFirstAtBeginDocument#1{#1}
\providecommand\HyField@AuxAddToFields[1]{}
\providecommand\HyField@AuxAddToCoFields[2]{}
\@writefile{toc}{\contentsline {chapter}{Introduction}{11}{chapter*.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {0.1}notation}{11}{section.0.1}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {1}Logistic Regression as a neural network}{13}{chapter.1}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {1.1}definitions}{13}{section.1.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.2}cost function}{13}{section.1.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.3}Gradient Descent}{14}{section.1.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.4}Model training}{15}{section.1.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.5}Forward Propagation}{16}{section.1.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {1.5.1}Activation Functions}{16}{subsection.1.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.6}Backward Propagation}{16}{section.1.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.7}Update parameters}{17}{section.1.7}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.8}Summary}{17}{section.1.8}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.9}Logistic Regression in Python}{17}{section.1.9}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {1.10}References}{18}{section.1.10}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {2}Neural Networks}{19}{chapter.2}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {2.1}Lingua franca}{19}{section.2.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.2}Model training}{20}{section.2.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.3}Parameter initialization}{21}{section.2.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {2.3.1}Xavier initialization}{22}{subsection.2.3.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.4}Forward Propagation}{22}{section.2.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.5}Backward Propagation}{22}{section.2.5}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.6}Summary}{23}{section.2.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {2.7}Deep neural networks in Python}{23}{section.2.7}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {3}Neural Networks hyperparameters}{25}{chapter.3}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.1}training}{25}{subsection.3.0.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.2}bias-variance trade-off}{26}{subsection.3.0.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.3}recipes for high-bias, high-variance}{26}{subsection.3.0.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.4}Regularization (weight decay)}{27}{subsection.3.0.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.5}Inverted Dropout Regularization}{28}{subsection.3.0.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.6}Input Normalization}{28}{subsection.3.0.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.7}Vanishing/Exploding gradients}{28}{subsection.3.0.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.0.8}gradient checking}{29}{subsection.3.0.8}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.1}Optimization}{29}{section.3.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.1.1}stochastic, mini-batch, and batch gradient descent}{30}{subsection.3.1.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.2}Gradient descent with momentum}{30}{section.3.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.3}RMSprob}{31}{section.3.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.4}Adaptive Momentum estimation (Adam)}{31}{section.3.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.4.1}Hyperparameters}{32}{subsection.3.4.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.5}Learning rate decay}{32}{section.3.5}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.6}Tuning parameters}{32}{section.3.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.6.1}coarse to fine search}{33}{subsection.3.6.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.6.2}panda vs caviar training approaches}{33}{subsection.3.6.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.7}Batch Normalization}{33}{section.3.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.7.1}Covariate Shift}{34}{subsection.3.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.7.2}Batch normalization as a regularization technique}{34}{subsection.3.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.7.3}Batch normalization on test sets}{35}{subsection.3.7.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.8}Multi-class classification}{35}{section.3.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.8.1}Softmax activation}{35}{subsection.3.8.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.9}Support Vector Machine (SVM)}{35}{section.3.9}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.9.1}Large Margins in linear Decision Boundaries}{38}{subsection.3.9.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.9.2}Non-linear Decision Boundaries}{38}{subsection.3.9.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.9.3}Gaussian kernel}{38}{subsection.3.9.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {3.10}Unsupervised Learning (Introduction)}{39}{section.3.10}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.10.1}k-mean algorithm}{39}{subsection.3.10.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {3.10.2}Principle Component Analysis PCA}{40}{subsection.3.10.2}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {4}Structuring machine learning}{43}{chapter.4}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {4.1}Machine learning strategy}{43}{section.4.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.2}orthogonality}{43}{section.4.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.3}Set up your Goal}{44}{section.4.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.3.1}Evaluation metric}{44}{subsection.4.3.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.3.2}Precision vs accuracy}{44}{subsection.4.3.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.3.3}F1-score}{45}{subsection.4.3.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.3.4}Satisfying-Optimizing metric}{45}{subsection.4.3.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.4}train/dev/test sets}{45}{section.4.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.4.1}dev/test metric}{46}{subsection.4.4.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.5}Human-level performance}{46}{section.4.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.5.1}human(bayes) error}{46}{subsection.4.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.6}Error analysis}{47}{section.4.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.6.1}Mislabeled data}{48}{subsection.4.6.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.7}Mismatched training dev/test sets}{48}{section.4.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.7.1}Addressing data mismatch}{49}{subsection.4.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.7.2}Artificial data synthesis}{49}{subsection.4.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.8}Multiple tasks learning}{50}{section.4.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.8.1}Transfer learning}{50}{subsection.4.8.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {4.8.2}Multitask learning}{50}{subsection.4.8.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {4.9}End-to-End Machine Learning}{50}{section.4.9}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {5}Computer Vision}{51}{chapter.5}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {5.1}Object Detection}{51}{section.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.1}Edge Detection}{52}{subsection.5.1.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.2}Padding}{53}{subsection.5.1.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.3}striding}{54}{subsection.5.1.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.4}Convolution on RGB channels}{54}{subsection.5.1.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.5}Multiple filters Convolution}{54}{subsection.5.1.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.6}Example ConvNet}{54}{subsection.5.1.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.7}Pooling Convolutions}{55}{subsection.5.1.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.8}Max Pooling}{55}{subsection.5.1.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.1.9}Average Pooling}{56}{subsection.5.1.9}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.2}Examples}{56}{section.5.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.2.1}LeNet-5 Network}{56}{subsection.5.2.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.2.2}AlexNet Network}{56}{subsection.5.2.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.2.3}VGG-16 Network}{57}{subsection.5.2.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.2.4}ResNet (Residual Block)}{58}{subsection.5.2.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.3}Inception}{58}{section.5.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.3.1}$1\times {1}$ Convolutions}{59}{subsection.5.3.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.3.2}Inception Block}{59}{subsection.5.3.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.3.3}Inception Branches}{59}{subsection.5.3.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.4}Object Detection}{59}{section.5.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.1}Localization and Detection}{59}{subsection.5.4.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.2}classification with localization}{60}{subsection.5.4.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.3}Landmark detection}{60}{subsection.5.4.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.4}Object Detection}{61}{subsection.5.4.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.5}Convolutional sliding windows}{61}{subsection.5.4.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.4.6}Example:}{61}{subsection.5.4.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.5}Bounding Box prediction}{62}{section.5.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.5.1}YOLO}{62}{subsection.5.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.5.2}Jaccard distance, or Intersection over Union IOU function}{62}{subsection.5.5.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.5.3}Non-Max Suppression}{63}{subsection.5.5.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.5.4}Anchor Boxes}{63}{subsection.5.5.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.5.5}Region Proposal R-CNN, Fast R-CNN}{63}{subsection.5.5.5}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.6}Face Recognition }{64}{section.5.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.1}Recognition vs. Verification}{64}{subsection.5.6.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.2}One-Shot learning}{64}{subsection.5.6.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.3}Similarity learning function}{64}{subsection.5.6.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.4}Siamese Network}{65}{subsection.5.6.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.5}Triplet Loss}{65}{subsection.5.6.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.6.6}Binary Classification}{66}{subsection.5.6.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {5.7}Neural Style Transfer}{67}{section.5.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.7.1}Generated image descent}{67}{subsection.5.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.7.2}Content cost}{67}{subsection.5.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.7.3}Style cost}{68}{subsection.5.7.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {5.7.4}Gram Matrix(style)}{68}{subsection.5.7.4}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {6}Recurrent Neural Networks RNN}{69}{chapter.6}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {6.1}Introduction}{69}{section.6.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.2}Notation}{69}{section.6.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.3}RNN}{70}{section.6.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.3.1}RNN model}{70}{subsection.6.3.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.4}uni-directional RNN Forward propagation}{70}{section.6.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.5}uni-directional RNN backward propagation through time}{71}{section.6.5}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.6}Variations of RNN models}{71}{section.6.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.6.1}many-to-one RNN}{72}{subsection.6.6.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.6.2}One-to-Many RNN}{72}{subsection.6.6.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.6.3}Many-to-Many of different input/output length}{74}{subsection.6.6.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.7}Language Model and sequence generation}{74}{section.6.7}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.8}Sampling novel sequences}{75}{section.6.8}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.9}Vanishing/Exploding gradients}{75}{section.6.9}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.9.1}Gated Recurrent Unit (GRU)}{75}{subsection.6.9.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.9.2}GRU simplified}{76}{subsection.6.9.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.9.3}Long Short Term Memory (LSTM)}{77}{subsection.6.9.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.9.4}Back-propagation}{78}{subsection.6.9.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.9.5}Bidirectional RNN}{78}{subsection.6.9.5}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.10}Deep RNNs}{78}{section.6.10}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.11}Word Representation}{79}{section.6.11}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.1}Word Embedding}{79}{subsection.6.11.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.2}Named Entity recognition}{79}{subsection.6.11.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.3}Learning word embedding}{80}{subsection.6.11.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.4}Word2Vec}{80}{subsection.6.11.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.5}Skip gram}{80}{subsection.6.11.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.6}hierarchical SOFTMAX}{80}{subsection.6.11.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.7}Negative Sampling}{81}{subsection.6.11.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.11.8}GloVe (global vector for word representation)}{82}{subsection.6.11.8}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.12}Example: Sentiment Classification}{83}{section.6.12}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.13}Sequence To Sequence Model}{83}{section.6.13}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.13.1}Beam search}{84}{subsection.6.13.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.13.2}normalization}{84}{subsection.6.13.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.13.3}Error Analysis}{85}{subsection.6.13.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.13.4}Bleu Score}{85}{subsection.6.13.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {6.14}Attention Model}{86}{section.6.14}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {6.14.1}Attention}{86}{subsection.6.14.1}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {7}Probabilistic Graphical Models (PGM)}{87}{chapter.7}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {7.1}Introduction}{87}{section.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.1.1}preliminaries}{88}{subsection.7.1.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.1.2}factors}{88}{subsection.7.1.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.2}Bayesian Network Fundamentals}{89}{section.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.2.1}constructing dependencies}{89}{subsection.7.2.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.3}Reasoning Patterns on BN}{90}{section.7.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.4}Flow of Probabilistic Influence}{91}{section.7.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.5}d-separation}{91}{section.7.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.5.1}I-map}{92}{subsection.7.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.6}Naive Bayes Model}{92}{section.7.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.7}Template Model}{93}{section.7.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.7.1}Distribution over Trajectories}{94}{subsection.7.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.7.2}Markov chain}{94}{subsection.7.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.7.3}Dynamic Bayesian Network DBN}{95}{subsection.7.7.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.7.4}Hidden Markov Model (HMM)}{95}{subsection.7.7.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {7.8}Structured CPD}{95}{section.7.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.8.1}Tree CPD}{96}{subsection.7.8.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {7.8.2}Multiplexer CPD}{97}{subsection.7.8.2}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {8}Natural language processing}{99}{chapter.8}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {8.1}pre-processing}{99}{section.8.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.2}Example: positive, negative classifier}{99}{section.8.2}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.3}Logistic regression classifier}{100}{section.8.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.4}Naive Bayes classifier}{101}{section.8.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.5}cosine similaritis}{103}{section.8.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.5.1}Euclidean distance}{104}{subsection.8.5.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.6}Principle Component Analysis (PCA)}{104}{section.8.6}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.7}Machine Translation}{105}{section.8.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.1}Loss function L}{106}{subsection.8.7.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.2}gradient descent}{107}{subsection.8.7.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.3}fixed number of iterations}{108}{subsection.8.7.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.4}k-Nearest neighbors algorithm}{108}{subsection.8.7.4}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.5}Searching for the translation embedding}{109}{subsection.8.7.5}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.6}LSH and document search}{109}{subsection.8.7.6}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.7}Bag-of-words (BOW) document models}{110}{subsection.8.7.7}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.8} Choosing the number of planes}{110}{subsection.8.7.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.7.9} Getting the hash number for a vector}{111}{subsection.8.7.9}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.8}Probabilistic model of pronounciation and spelling}{112}{section.8.8}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.8.1}auto-correction}{112}{subsection.8.8.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.8.2}Bayesian inference model}{113}{subsection.8.8.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.8.3}Minimum edit distance}{114}{subsection.8.8.3}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.9}Grammar Weighted Automata}{114}{section.8.9}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.9.1}Markov chain}{114}{subsection.8.9.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.9.2}Hidden Markov Models HMMs}{115}{subsection.8.9.2}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.9.3}Viterbi Algorithm}{116}{subsection.8.9.3}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.9.4}Part of Speech tagging (POS)}{117}{subsection.8.9.4}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {8.10}N-grams}{118}{section.8.10}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.10.1}Smoothing}{118}{subsection.8.10.1}\protected@file@percent }
\@writefile{toc}{\contentsline {subsection}{\numberline {8.10.2}Back-off}{118}{subsection.8.10.2}\protected@file@percent }
\citation{xyz2}
\bibstyle{apalike}
\bibdata{bib0}
\@writefile{toc}{\contentsline {subsection}{\numberline {8.10.3}Interpolation}{119}{subsection.8.10.3}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{Appendices}{121}{section*.4}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {A}Introduction to probabilities}{123}{appendix.a.A}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {section}{\numberline {A.1}probabilities chain rule}{123}{section.a.A.1}\protected@file@percent }
\@writefile{toc}{\contentsline {section}{\numberline {A.2}Naive Bayes}{123}{section.a.A.2}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {B}Covariance}{125}{appendix.a.B}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {chapter}{\numberline {C}Single Value Decomposition}{127}{appendix.a.C}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {chapter}{\numberline {D}Exponentially weighted averages}{129}{appendix.a.D}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {subsection}{\numberline {D.0.1}How to choose the value $\beta $ ?}{129}{subsection.a.D.0.1}\protected@file@percent }
\@writefile{toc}{\contentsline {chapter}{\numberline {E}smoothing (add-k, and add-one Laplacian)}{131}{appendix.a.E}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\contentsline {chapter}{\numberline {F}Kernels, and Convolution functions}{133}{appendix.a.F}\protected@file@percent }
\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\addvspace {10\p@ }}