Scribd
Upload
150 views39 pages

Deep Learning Techniques Overview

This document provides an overview of deep learning techniques. It includes 15 chapters that cover topics such as pattern recognition, supervised learning, neural networks, training neural networks, convolutional neural networks, recurrent neural networks, and deep learning applications in R and Python. Code examples in Python demonstrate how to build and train neural network models on classification tasks.

Uploaded by

william
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
150 views39 pages

Deep Learning Techniques Overview

This document provides an overview of deep learning techniques. It includes 15 chapters that cover topics such as pattern recognition, supervised learning, neural networks, training neural networks, convolutional neural networks, recurrent neural networks, and deep learning applications in R and Python. Code examples in Python demonstrate how to build and train neural network models on classification tasks.

Uploaded by

william
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
  • Preface
  • Introduction
  • Pattern Recognition
  • Supervised Learning
  • Linear Learning Models
  • Neural Networks with Deep Learning
  • Training Neural Networks
  • Techniques to Improve Gradient Descent
  • Techniques to Improve Model Generalization Ability
  • Hyper-Parameter Selection
  • Deep Learning with R
  • Convolutional Neural Networks
  • Deep Learning with Python
  • Recurrent Neural Networks
  • Glossary
  • Placeholder
  • Bibliography

Deep Learning

Subir Varma and Sanjiv Das


2017-10-17
2
Contents

Preface 5

1 Introduction 7

2 Pattern Recognition 9

3 Supervised Learning 11

4 Linear Learning Models 13

5 Neural Networks with Deep Learning 15

6 Training Neural Networks 17

7 Techniques to Improve Gradient Descent 19

8 Techniques to Improve Model Generalization Ability 21

9 Hyper-Parameter Selection 23

10 Deep Learning with R 25

11 Convolutional Neural Networks 27

12 Deep Learning with Python 29

13 Recurrent Neural Networks 33

14 Glossary 35

15 Placeholder 37

3
4 CONTENTS
Preface

5
6 CONTENTS
Chapter 1

Introduction

7
8 CHAPTER 1. INTRODUCTION
Chapter 2

Pattern Recognition

9
10 CHAPTER 2. PATTERN RECOGNITION
Chapter 3

Supervised Learning

11
12 CHAPTER 3. SUPERVISED LEARNING
Chapter 4

Linear Learning Models

13
14 CHAPTER 4. LINEAR LEARNING MODELS
Chapter 5

Neural Networks with Deep Learning

15
16 CHAPTER 5. NEURAL NETWORKS WITH DEEP LEARNING
Chapter 6

Training Neural Networks

17
18 CHAPTER 6. TRAINING NEURAL NETWORKS
Chapter 7

Techniques to Improve Gradient


Descent

19
20 CHAPTER 7. TECHNIQUES TO IMPROVE GRADIENT DESCENT
Chapter 8

Techniques to Improve Model


Generalization Ability

21
22 CHAPTER 8. TECHNIQUES TO IMPROVE MODEL GENERALIZATION ABILITY
Chapter 9

Hyper-Parameter Selection

23
24 CHAPTER 9. HYPER-PARAMETER SELECTION
Chapter 10

Deep Learning with R

25
26 CHAPTER 10. DEEP LEARNING WITH R
Chapter 11

Convolutional Neural Networks

27
28 CHAPTER 11. CONVOLUTIONAL NEURAL NETWORKS
Chapter 12

Deep Learning with Python

import pylab

from [Link] import Sequential


from [Link] import Dense, Dropout, Activation
from [Link].advanced_activations import LeakyReLU
from keras import backend
from [Link] import to_categorical

#Read in data
import pandas as pd
tf_train = pd.read_csv("[Link]")
tf_test = pd.read_csv("[Link]")
n = len(tf_train)

X_train = tf_train.iloc[:,1:10].values
X_test = tf_test.iloc[:,1:10].values
y_train = tf_train.iloc[:,10].values
y_test = tf_test.iloc[:,10].values

idx = y_train; y_train = zeros(len(idx));


y_train[idx=='malignant']=1; y_train = y_train.astype(int)

idx = y_test; y_test = zeros(len(idx));


y_test[idx=='malignant']=1; y_test = y_test.astype(int)

Y_train = to_categorical(y_train,2)

#Set up and compile the model


model = Sequential()
n_units = 512
data_dim = X_train.shape[1]

[Link](Dense(n_units, input_dim=data_dim))
[Link](LeakyReLU())
[Link](Dropout(0.25))

[Link](Dense(n_units))

29
30 CHAPTER 12. DEEP LEARNING WITH PYTHON

Figure 12.1: Loss functions and accuracy for the training and validation data sets

[Link](LeakyReLU())
[Link](Dropout(0.25))

[Link](Dense(n_units))
[Link](LeakyReLU())
[Link](Dropout(0.25))

[Link](Dense(n_units))
[Link](LeakyReLU())
[Link](Dropout(0.25))

[Link](Dense(n_units))
[Link](LeakyReLU())
[Link](Dropout(0.25))

[Link](Dense(2))
[Link](Activation('sigmoid'))

[Link](optimizer='rmsprop',
loss='binary_crossentropy',
metrics=['accuracy'])

#Fit the model


bsize = 32
[Link](X_train, Y_train, batch_size=bsize, epochs=15, validation_split=0.1, verbose=2)

## OUT-SAMPLE ACCURACY
from [Link] import confusion_matrix
yhat = model.predict_classes(X_test,batch_size=bsize)
cm = confusion_matrix(y_test,yhat)
print("Confusion matrix = "); print(cm)
31

acc = sum(diag(cm))/sum(cm)
print("Accuracy = ",acc)
32 CHAPTER 12. DEEP LEARNING WITH PYTHON
Chapter 13

Recurrent Neural Networks

33
34 CHAPTER 13. RECURRENT NEURAL NETWORKS
Chapter 14

Glossary

35
36 CHAPTER 14. GLOSSARY
Chapter 15

Placeholder

37
38 CHAPTER 15. PLACEHOLDER
Bibliography

39

Deep Learning Subir Varma and Sanjiv Das 2017-10-17
2
Contents Preface 5 1 Introduction 7 2 Pattern Recognition 9 3 Supervised Learning 11 4 Linear Learning Models 13 5 Neural Net
4 CONTENTS
Preface 5
6 CONTENTS
Chapter 1 Introduction 7
8 CHAPTER 1. INTRODUCTION
Chapter 2 Pattern Recognition 9
10 CHAPTER 2. PATTERN RECOGNITION

You might also like