Preface
Acknowledgements
Contents
List of Figures
List of Tables
Part I: Introduction and Overview
Chapter 1: Introduction
1.1 Artificial Intelligence, Machine Learning, Deep Learning
1.1.1 Artificial Intelligence
1.1.2 Machine Learning
1.1.3 Deep Learning
1.1.4 How they Come Together
1.2 Artificial Intelligence Is Driving Innovation
1.2.1 Transforming Healthcare
1.2.2 Protecting Wildlife
1.2.3 Securing the Environment
1.3 Tools, Frameworks and Hardware
1.3.1 Building Intelligent Applications
1.3.2 Python, Notebooks and Environments
1.3.3 Pre-Processing
1.3.4 Machine Learning
1.3.5 Deep Learning
1.3.6 Inferencing
1.4 How this Book Is Organised
1.5 Who Should Read this Book
1.6 Summary
References
Part II: Foundations of Machine Learning
Chapter 2: Fundamentals of Machine Learning
2.1 What Is Machine Learning?
2.1.1 Formal and Non-Formal Definition
2.1.2 How AI and Machine Learning Differs from Conventional Software Development
2.1.2.1 Rewriting the Rules
2.1.2.2 Intelligent Decision Making
2.2 Machine Learning Tribes
2.2.1 Connectionists
2.2.2 Evolutionists
2.2.3 Bayesians
2.2.4 Symbolists
2.2.5 Analogists
2.3 Data Management
2.3.1 Data Types and Data Objects
2.3.1.1 Numerical
2.3.1.2 Textual
2.3.1.3 Categorical
2.3.1.4 Timeseries
2.3.2 Data Structure
2.3.2.1 Data Objects
2.3.3 Datasets
2.3.4 Exploratory Data Analysis
2.3.4.1 What Is Exploratory Data Analysis
2.3.4.2 Data Distributions
2.3.4.3 Validate Assumptions
2.3.4.4 Feature Engineering
2.4 Learning Problems
2.4.1 Supervised Machine Learning
2.4.2 Semi-Supervised Machine Learning
2.4.3 Un-Supervised Machine Learning
2.4.4 Regression
2.4.5 Reinforcement Learning
2.5 Evaluating Machine Learning Models
2.6 Summary
References
Chapter 3: Supervised Learning
3.1 Basic Concepts
3.2 Supervised Learning Tasks
3.2.1 Data Extraction
3.2.2 Data Preparation
3.2.2.1 Data Size
3.2.2.2 Missing Data
3.2.2.3 Textual Data
One Hot Encoding
3.2.2.4 Value Ranges (Normalisation and Scaling)
Scaling
Normalisation
Standardisation
3.2.2.5 Distribution
3.2.2.6 Class Balance
3.2.2.7 Correlation Between Features
3.2.3 Feature Engineering
3.2.3.1 Feature Selection
3.2.3.2 Dimensionality Reduction
3.2.4 Selecting a Training Algorithm
3.3 Supervised Algorithms
3.3.1 Linear Regression
3.3.2 Logistic Regression
3.3.3 Linear Discriminate Analysis
3.3.4 Support Vector Machine
3.3.5 Random Forest
3.3.6 Naive Bayes
3.3.7 K-Nearest Neighbours
3.4 Summary
References
Chapter 4: Un-Supervised Learning
4.1 Basic Concepts
4.2 Clustering
4.2.1 Hierarchical Clustering
4.2.2 K-Means
4.2.3 Mixture Models
4.2.4 DBSCAN
4.2.5 Optics Algorithm
4.3 Principal Component Analysis
4.4 Association Rule Mining
4.5 Summary
References
Chapter 5: Performance Evaluation Metrics
5.1 Introduction to Model Evaluation
5.1.1 Evaluation Challenges
5.1.2 Taxonomy of Classifier Evaluation Metrics
5.2 Classification Accuracy
5.3 Train, Test and Validation Sets
5.4 Underfitting and Overfitting
5.5 Supervised Learning Evaluation Metrics
5.5.1 Confusion Matrix
5.5.1.1 Accuracy
5.5.1.2 Precision
5.5.1.3 Recall (Sensitivity)
5.5.1.4 Specificity
5.5.1.5 False Positive Rate
5.5.1.6 F1-Score
5.5.2 Receiver Operating Characteristic
5.5.3 Regression Metrics
5.5.3.1 Mean Square Error (MSE)
5.5.3.2 MAE
5.5.3.3 R2 (Coefficient of Determination)
5.6 Probability Scoring Methods
5.6.1 Log Loss Score
5.6.2 Brier Score
5.7 Cross-Validation
5.7.1 Challenge of Evaluating Classifiers
5.7.2 K-Fold Cross-Validation
5.8 Un-Supervised Learning Evaluation Metric
5.8.1 Elbow Method
5.8.2 Davies-Bouldin Index
5.8.3 Dunn Index
5.8.4 Silhouette Coefficient
5.9 Summary
References
Part III: Deep Learning Concepts and Techniques
Chapter 6: Introduction to Deep Learning
6.1 So whatΒ΄s the Difference Between DL and ML?
6.2 Introduction to Deep Learning
6.3 Artificial Neural Networks
6.3.1 Perceptrons
6.3.2 Neural Networks
6.3.3 Activation Functions
6.3.4 Multi-Class Classification Considerations
6.3.5 Cost Functions and Optimisers
6.3.6 Backpropagation
6.3.7 The Vanishing Gradient
6.3.8 Weight Initialisation
6.3.9 Regularisation
6.4 Convolutional Neural Networks
6.4.1 Image Filters and Kernels
6.4.2 Convolutional Layers
6.4.3 Pooling Layers
6.4.4 Transfer Learning
6.5 Summary
References
Chapter 7: Image Classification and Object Detection
7.1 Hardware Accelerated Deep Learning
7.1.1 Training and Associated Hardware
7.1.1.1 Development Systems
7.1.1.2 Training Systems
7.1.1.3 Inferencing Systems
7.1.2 Tensor Processing Unit (TPU)
7.1.3 Other Hardware Considerations
7.2 Object Recognition
7.2.1 Image Classification
7.2.2 Object Detection
7.2.3 Semantic Segmentation
7.2.4 Object Segmentation
7.3 Model Architectures
7.3.1 Single Shot Detector (SSD)
7.3.2 YOLO Family
7.3.3 R-CNN
7.3.4 Fast-RCNN
7.3.5 Faster-RCNN
7.3.6 EfficientNet
7.3.7 Comparing Architectures
7.3.7.1 Key Findings
7.3.7.2 Most Accurate
7.3.7.3 Fastest
7.4 Evaluation Metrics
7.4.1 Confidence Score
7.4.2 Intersection over Union
7.4.3 Mean Average Precision (mAP)
7.5 Summary
References
Chapter 8: Deep Learning Techniques for Time Series Modelling
8.1 Introduction to Time-Series Data
8.2 Recurrent Neural Network
8.2.1 Developing RNNs for Time Series Forecasting
8.3 Long-Term Short-Term Memory
8.4 Gated Recurrent Unit
8.5 One Dimensional Convolutional Neural Network
8.6 Summary
References
Chapter 9: Natural Language Processing
9.1 Introduction to Natural Language Processing
9.1.1 Tokenisation
9.1.2 Stemming
9.1.3 Lemmatization
9.1.4 Stop Words
9.1.5 Phrase Matching and Vocabulary
9.2 Text Classification
9.2.1 Text Feature Extraction
9.3 Sentiment Analysis
9.4 Topic Modelling
9.4.1 Latent Semantic Analysis (LSA)
9.4.2 Latent Dirichlet Allocation
9.4.3 Non-negative Matrix Factorization
9.5 Deep Learning for NLP
9.5.1 Word Embeddings
9.5.2 Word Embedding Algorithms
9.5.2.1 Embedding Layer
9.5.2.2 Word2Vec
9.5.2.3 GloVe
9.5.2.4 Natural Language Understanding and Generation
9.6 Real-World Applications
9.6.1 Chat Bots
9.6.2 Smart Speakers
9.7 Summary
References
Chapter 10: Deep Generative Models
10.1 Autoencoders
10.1.1 Autoencoder Basics
10.1.2 Autoencoder for Dimensionality Reduction
10.1.3 Autoencoder for Images
10.1.4 Stacked Autoencoders
10.1.5 Generative Adversarial Networks (GANS)
10.1.5.1 GANs Network Architectures
10.2 Summary
References
Chapter 11: Deep Reinforcement Learning
11.1 What Is Reinforcement Learning?
11.2 Reinforcement Learning Definitions
11.3 Domain Selection for Reinforcement Learning
11.4 State-Action Pairs and Complex Probability Distributions of Reward
11.5 Neural Networks and Reinforcement Learning
11.6 The Deep Reinforcement Learning Process
11.7 Practical Applications of Deep Reinforcement Learning
11.8 Summary
References
Part IV: Enterprise Machine Learning
Chapter 12: Accelerated Machine Learning
12.1 Introduction
12.1.1 CPU/GPU Based Clusters
12.2 CPU Accelerated Computing
12.2.1 Distributed Accelerated Computing Frameworks
12.2.1.1 Local Vs Distributed
12.2.1.2 Benefits of Scaling Out
12.2.1.3 Hadoop
12.2.1.4 Apache Spark
12.3 Introduction to DASK
12.3.1 DASK Arrays
12.3.2 Scikit Learn and DASK Integration (DASK ML)
12.3.3 Scikit Learn Joblib
12.4 GPU Computing
12.4.1 Introduction to GPU Hardware
12.4.2 Introduction to NVIDIA Accelerated Computing
12.4.3 CUDA
12.4.4 CUDA Accelerated Computing Libraries
12.5 RAPIDS
12.5.1 cuDF Analytics
12.5.2 cuML Machine Learning
12.5.3 cuGraph Graph Analytics
12.5.4 Apache Arrow
12.6 Summary
References
Chapter 13: Deploying and Hosting Machine Learning Models
13.1 Introduction to Deployment
13.1.1 Why Is Model Deployment Important
13.1.2 Enabling MLOps
13.1.3 MLOps Frameworks
13.1.4 MLOps Application Programmable Interfaces APIΒ΄s
13.2 Preparing a Model
13.2.1 Model Formats
13.2.1.1 ProtoBuf (pb)
13.2.1.2 ONNX (.ONNX)
13.2.1.3 Keras h5 (.h5)
13.2.1.4 TensorFlow SavedModel Format
13.2.1.5 Scikit-Learn (.pkl)
13.2.1.6 IOS Platform (.mlmodel)
13.2.1.7 PyTorch (.pt)
13.2.2 Freezing and Exporting Models
13.2.3 Model Optimisation
13.2.4 Deploying the TFLite Model and Undertaking Inference
13.3 Web Deployment
13.3.1 Flask
13.3.2 Why Use Flask
13.3.3 Working and Developing in Flask
13.4 Summary
References
Chapter 14: Enterprise Machine Learning Serving
14.1 Docker
14.1.1 What Is Docker
14.1.2 Working with Docker
14.1.2.1 Using Docker
14.1.2.2 WhatΒ΄s a Container
14.1.2.3 Docker Run
14.1.2.4 Container Lifecycle
14.1.2.5 Building Custom Dockers
14.1.3 Docker Compose
14.1.4 Docker Volume and Mount
14.2 Kubernetes
14.3 TensorFlow Serving
14.3.1 Why Use TensorFlow Serving
14.3.2 TensorFlow Serving on CPU and GPU (NVidia Runtime)
14.4 Summary
References