Machine Learning for Computer Scientists and Data Analysts: From an Applied Perspective

Preface
Contents
Part I Basics of Machine Learning
	1 What Is Applied Machine Learning?
		1.1 Introduction
		1.2 The Machine Learning Pipeline
		1.3 Knowing the Application and Data
		1.4 Getting Started Using Python
		1.5 Metadata Extraction and Data Pre-processing
		1.6 Data Exploration
		1.7 A Practice for Performing Exploratory Data Analysis
			1.7.1 Importing the Required Libraries for EDA
			1.7.2 Loading the Data Into Dataframe
			1.7.3 Data Visualization
				2D Scatter Plot
				Pair-Plot
				Histogram Plot
				Probability Distribution Function
				Cumulative Distribution Function
				Box Plot
				Violin Plots
				Univariate, Bivariate, and Multivariate Analysis
			1.7.4 Data Analysis
				Standard Deviation
				Mean/Average
				Variance
			Median
				Percentile
				Quantile
				Interquartile Range
				Mean Absolute Deviation
			1.7.5 Performance Evaluation Metrics
		1.8 Putting It All Together
		1.9 Exercise Problems
	2 A Brief Review of Probability Theory and Linear Algebra
		2.1 Introduction
		2.2 Fundamental of the Probability
			Independence
		2.3 Discrete Random Variable
			2.3.1 Probability Mass Function
			2.3.2 Cumulative Distribution Function
			2.3.3 Expectation and Variance
				Expectation
				Variance
		2.4 Continuous Random Variable
			2.4.1 Probability Density Function
			2.4.2 Expectation and Variance
				Expectation
				Variance
		2.5 Common Distributions
			2.5.1 Discrete Distributions
				Bernoulli Distribution
				Binomial Distribution
				The Multinomial Distributions
				The Poisson Distribution
			2.5.2 Continuous Distributions
				Gaussian (Normal) Distribution
				Exponential Distribution
				Memory-Less Property of Exponential Distribution
				Uniform Distribution
		2.6 Joint Probability Distributions
			2.6.1 Joint Distribution: Discrete Random Variables
				Joint CDF
				Marginal PMF
			2.6.2 Joint Distribution: Continuous Random Variables
				Joint CDF
				Marginal PDF
			2.6.3 Covariance and Correlation
				Covariance
				Correlation
			2.6.4 Multivariate Gaussian Distribution
		2.7 Matrix Decomposition
			2.7.1 Eigenvalue Decomposition
				Eigendecomposition
			2.7.2 Singular Value Decomposition
		2.8 Putting It All Together
		2.9 Exercise Problems
	3 Supervised Learning
		3.1 Introduction
		3.2 Preparing Data
			3.2.1 Data Abstraction
			3.2.2 Dealing with Missing Data
			3.2.3 Dealing with Imbalanced Datasets
		3.3 Regression
			3.3.1 Linear Regression
			3.3.2 Multi-Variable Linear Regression
			3.3.3 Multi-Variable Adaptive Regression Splines (MARS)
			3.3.4 AutoRegressive Moving Average
			3.3.5 Bayesian Linear Regression
			3.3.6 Logistic Regression
		3.4 Artificial Neural Networks
			3.4.1 Modeling of Neuron
			3.4.2 Implementing Logical Gates with ANN
			3.4.3 Multi-Layer Perceptron
				Activation Functions
			3.4.4 Training of MLPs
				Stochastic Gradient Descent (SGD)
				Mini-Batch Gradient Descent
				Adagrad
				Adam
			3.4.5 Inference
			3.4.6 Issues with Multi-Layer Perceptron
				Overfitting
				Underfitting
				Overfitting vs Underfitting
			3.4.7 Instances of Deep Neural Networks
				Convolutional Neural Networks
				Radial Basis Function (RBF) Neural Networks
				Recurrent Neural Networks
				Long-Short-Term Memory Neural Networks
		3.5 Support Vector Machines
			3.5.1 SVM Kernels
			3.5.2 Multiclass Classification
		3.6 Ensemble Learning
			3.6.1 Bagging
			3.6.2 AdaBoost
			3.6.3 Bootstrap
			3.6.4 Gradient Boosting
			3.6.5 Stacking
		3.7 Other Machine Learning Techniques
			3.7.1 Bayesian Model Combination
			3.7.2 Random Forest
			3.7.3 Tree-Based Methods
			3.7.4 AutoEncoder
				Stacked Autoencoders
				Convolutional Autoencoders
				Sparse Autoencoders
				Recurrent Autoencoders
				Denoising Autoencoders
				Variational Autoencoders (VAE)
		3.8 Putting It All Together
		3.9 Exercise Problems
	4 Unsupervised Learning
		4.1 Introduction
		4.2 Clustering
			4.2.1 K-Means Clustering
			4.2.2 Hierarchical Clustering
			4.2.3 Mixture Models
		4.3 Unsupervised Neural Networks
			4.3.1 Self-Organizing Maps
			4.3.2 Generative Adversarial Networks
			4.3.3 Deep Belief Nets
				Training a DBN
				Applying the DBN
			4.3.4 Method of Moments
		4.4 Feature Selection Techniques
			4.4.1 Principal Component Analysis
			4.4.2 T-Distributed Stochastic Neighbor Embedding
			4.4.3 Pearson Correlation Coefficient
			4.4.4 Independent Component Analysis
			4.4.5 Non-negative Matrix Factorization (NMF)
		4.5 Multi-Dimensional Scaling
		4.6 Google Page Ranking Algorithm
		4.7 Putting It All Together
		4.8 Exercise Problems
	5 Reinforcement Learning
		5.1 Introduction
		5.2 Q-Learning
			5.2.1 Accelerated Q-learning by Environment Exploration
		5.3 TD(λ)-Learning
		5.4 SARSA Learning
		5.5 Deep Q-Learning
		5.6 Policy Optimization
			5.6.1 Stochastic Policy Gradient
			5.6.2 REINFORCE
				Action Value Expression
				Other Expressions
		5.7 Gradient-Based Policy Optimization
		5.8 Putting It All Together
		5.9 Exercise Problems
Part II Advanced Machine Learning
	6 Online Learning
		6.1 Introduction
		6.2 Online Supervised Learning
			6.2.1 First-/Second-Order Online Learning
				Passive Aggressive Online Learning (PA)
				Online Gradient Descent (OGD)
				Second-Order Perceptron (SOP)
			6.2.2 Online Learning with Regularization
				Truncated Gradient Descent
				Forward-Looking Subgradients (FOBOS)
				Regularized Dual Averaging (RDA)
				Follow-the-Regularized-Leader-Proximal (FTRL-Proximal)
		6.3 Online Unsupervised Learning
			6.3.1 Online Clustering
				Partition-Based Online Clustering
				Density-Based Online Clustering
			6.3.2 Other Unsupervised Tasks
				Online Dimension Reduction
				Online Density Estimation
				Online Anomaly Detection
		6.4 Application and Resources
			6.4.1 Time Series Prediction
			6.4.2 Information Retrieval
			6.4.3 Online Portfolio Selection
			6.4.4 Other Applications: Combined with Deep Learning
			6.4.5 Resources
		6.5 Putting It All Together
		6.6 Exercise Problems
	7 Recommender Learning
		7.1 Introduction
		7.2 The Recommendation Problem
		7.3 Content-Based Approach
		7.4 Collaborative Filtering
			7.4.1 Memory-Based Collaborative Filtering
				User Similarity
				Rating Aggregation
			7.4.2 Latent Factor Model
				Matrix Factorization as Collaborative Filtering
				From Matrix Factorization to Latent Factor Model
				Incorporating Rating Bias
				Regularization
		7.5 Factorization Machine
		7.6 Deep Learning Models
		7.7 Application and Resources
			7.7.1 Applications
			7.7.2 Resources
		7.8 Putting It All Together
		7.9 Exercise Problems
	8 Graph Learning
		8.1 Introduction
		8.2 Basics of Math
			8.2.1 Matrix Manipulation
			8.2.2 Eigendecomposition on Matrix
			8.2.3 Approximation Theory
			8.2.4 Graph Representations and Graph Signal
			8.2.5 Spectral Graph Theory
		8.3 Graph Neural Network Models
			8.3.1 Spatial-Based Graph Convolution Networks
				Linear Aggregation Function
				Polynomial Aggregation Function
				Rational Propagation Function
			8.3.2 Spectral-Based Graph Convolution Networks
				Linear Filter Function
				Polynomial Filter Function
				Rational Filter Function
			8.3.3 Other Graph Neural Networks
		8.4 Application and Resources
		8.5 Put It All Together
		8.6 Exercise Problems
	9 Adversarial Machine Learning
		9.1 Introduction
		9.2 Adversarial Attacks and Defenses
			9.2.1 Adversarial Attacks
				Fast Gradient Sign Method (FGSM)
				Basic Iterative Method (BIM)
				Momentum Iterative Method (MIM)
				Projected Gradient Descent Attack
				Jacobian-Based Saliency Map Attack (JSMA)
				DeepFool Attack
				Carlini and Wagner Attack (CW)
				One-Pixel Attack
				Universal Perturbation
			9.2.2 Adversarial Defenses
				Adversarial Training
				Defensive Distillation
				Random Ensemble
				Gradient Regularization
				MagNet
				Detecting Adversaries
				Autoencoders
		9.3 Experimental Results
			9.3.1 Network Architecture
			9.3.2 Performance with Adversarial Attacks
			9.3.3 Effective Adversarial Training
				Performance Evaluation and Comparison
		9.4 Putting It All Together
		9.5 Exercise Problems
Part III Machine Learning in the Field
	10 SensorNet: An Educational Neural Network Framework for Low-Power Multimodal Data Classification
		10.1 Introduction
		10.2 SensorNet Architecture
			10.2.1 Deep Neural Networks Overview
				Convolutional Layers
				Pooling Layers
				Fully Connected Layers
				Activation Functions
			10.2.2 Signal Preprocessing
			10.2.3 Neural Network Architecture
		10.3 SensorNet Evaluation using Three Case Studies
			10.3.1 Case Study 1: Physical Activity Monitoring
				Dataset
				Experiment Setup and Results
			10.3.2 Case Study 2: Stand-Alone Dual-Mode Tongue Drive System (sdTDS)
				sdTDS Overview and Experimental Setup
				Experiment Results
			10.3.3 Case Study 3: Stress Detection
				Dataset
				Experiment Setup and Results
		10.4 SensorNet Optimization and Complexity Reduction
			10.4.1 The Number of Convolutional Layers
			10.4.2 The Number of Filters
			10.4.3 Filter Shapes
			10.4.4 Zero-Padding
			10.4.5 Activation Functions
		10.5 SensorNet Hardware Architecture Design
			10.5.1 Exploiting Efficient Parallelism
			10.5.2 Hardware Performance Parameters
		10.6 Resources
		10.7 Exercise Problems
	11 Transfer Learning in Mobile Health
		11.1 Introduction
		11.2 Transfer Learning
		11.3 Problem Statement
			11.3.1 Problem Definition
			11.3.2 Problem Formulation
		11.4 TransFall Framework Design
			11.4.1 Vertical Transformation
			11.4.2 Horizontal Transformation
			11.4.3 Label Estimation
		11.5 Validation Approach
			11.5.1 Overview of the Datasets
			11.5.2 Cross-Domain Transfer Learning Scenarios
			11.5.3 Comparison Approach and Performance Metrics
			11.5.4 Choice of Classification Model
		11.6 Results
			11.6.1 Cross-Platform Transfer Learning Results
			11.6.2 Cross-Subject Transfer Learning Results
			11.6.3 Hybrid Transfer Learning Results
			11.6.4 Transformation Module Analysis
			11.6.5 Parameter Examination
		11.7 Exercise Problems
	12 Applied Machine Learning for Computer Architecture Security
		12.1 Introduction
			12.1.1 Malware
			12.1.2 Microarchitectural Side-Channel Attacks
		12.2 Challenges Associated with Traditional Security Mechanisms
		12.3 Deployment of Hardware Performance Counters for Computer Architecture Security
		12.4 Application of Machine Learning for Computer Architecture Security Countermeasures
			12.4.1 Feature Selection: Key Microarchitectural Features
		12.5 ML for Hardware-Assisted Malware Detection: Comparative Analysis
			12.5.1 Experimental Setup and Data Collection
			12.5.2 Feature Selection and ML Classifiers Implementation
			12.5.3 Evaluation Results of ML-Based Malware Detectors
		12.6 ML for Microarchitectural SCAs Detection: Comparative Analysis
			12.6.1 Detection Based on Victim Applications\' HPCs Data
			12.6.2 ML Classifiers Implementation
			12.6.3 Evaluation Results of ML-Based SCAs Detectors
		12.7 Exercise Problems
	13 Applied Machine Learning for Cloud Resource Management
		13.1 Introduction
			13.1.1 Challenge of Diversity
		13.2 Modern Resource Provisioning Systems: ML Comes to the Rescue
		13.3 Applications of Machine Learning in Resource Provisioning Systems
			13.3.1 Monitoring and Prediction of Applications\' Behavior
				Online Monitoring
				Phase Prediction
			13.3.2 Using ML for Performance/Cost/Energy Estimation
			13.3.3 Explore and Optimize the Selection
			13.3.4 Decision Making
		13.4 Security Threats in Cloud Rooted from ML-Based RPS
			13.4.1 Adversarial Machine Learning Attack to RPS
				Finding Physical Hosts Running Victim Instances
				Evasion from Detection and Migration
				Overview of Attack
			13.4.2 Isolation as a Remedy
		13.5 Exercise Problems
References
Index