Artificial Intelligence By Example: Acquire Advanced AI, Machine Learning and Deep Learning design skills. 2 Ed

Cover....1

Copyright....3

Packt Page....4

Contributors....5

Table of Contents....8

Preface....20

Chapter 1: Getting Started with Next-Generation Artificial Intelligence through Reinforcement Learning....30

Reinforcement learning concepts....31

How to adapt to machine thinking and become an adaptive thinker....33

Overcoming real-life issues using the three-step approach....34

Step 1 – describing a problem to solve: MDP in natural language....36

Watching the MDP agent at work....37

Step 2 – building a mathematical model: the mathematical representation of the Bellman equation and MDP....39

From MDP to the Bellman equation....39

Step 3 – writing source code: implementing the solution in Python....43

The lessons of reinforcement learning....45

How to use the outputs....47

Possible use cases....49

Machine learning versus traditional applications....52

Summary....53

Questions....53

Further reading....54

Chapter 2: Building a Reward Matrix – Designing Your Datasets....56

Designing datasets – where the dream stops and the hard work begins....57

Designing datasets....58

Using the McCulloch-Pitts neuron....58

The McCulloch-Pitts neuron....60

The Python-TensorFlow architecture....64

Logistic activation functions and classifiers....64

Overall architecture....64

Logistic classifier....65

Logistic function....66

Softmax....67

Summary....71

Questions....72

Further reading....72

Chapter 3: Machine Intelligence – Evaluation Functions and Numerical Convergence....74

Tracking down what to measure and deciding how to measure it....75

Convergence....77

Implicit convergence....78

Numerically controlled gradient descent convergence....78

Evaluating beyond human analytic capacity....85

Using supervised learning to evaluate a result that surpasses human analytic capacity....89

Summary....93

Questions....94

Further reading....94

Chapter 4: Optimizing Your Solutions with K-Means Clustering....96

Dataset optimization and control....97

Designing a dataset and choosing an MLDL model....98

Approval of the design matrix....99

Implementing a k-means clustering solution....103

The vision....103

The data....104

The strategy....105

The k-means clustering program....106

The mathematical definition of k-means clustering....107

The Python program....109

Saving and loading the model....113

Analyzing the results....114

Bot virtual clusters as a solution....115

The limits of the implementation of the k-means clustering algorithm....116

Summary....117

Questions....117

Further reading....118

Chapter 5: How to Use Decision Trees to Enhance K-Means Clustering....120

Unsupervised learning with KMC with large datasets....121

Identifying the difficulty of the problem....123

NP-hard – the meaning of P....123

NP-hard – the meaning of non-deterministic....124

Implementing random sampling with mini-batches....124

Using the LLN....125

The CLT....125

Using a Monte Carlo estimator....126

Trying to train the full training dataset....127

Training a random sample of the training dataset....127

Shuffling as another way to perform random sampling....129

Chaining supervised learning to verify unsupervised learning....131

Preprocessing raw data....132

A pipeline of scripts and ML algorithms....132

Step 1 – training and exporting data from an unsupervised ML algorithm....134

Step 2 – training a decision tree....135

Step 3 – a continuous cycle of KMC chained to a decision tree....139

Random forests as an alternative to decision trees....143

Summary....147

Questions....147

Further reading....148

Chapter 6: Innovating AI with Google Translate....150

Understanding innovation and disruption in AI....152

Is AI disruptive?....152

AI is based on mathematical theories that are not new....153

Neural networks are not new....153

Looking at disruption – the factors that are making AI disruptive....154

Cloud server power, data volumes, and web sharing of the early 21st century....154

Public awareness....155

Inventions versus innovations....155

Revolutionary versus disruptive solutions....156

Where to start?....156

Discover a world of opportunities with Google Translate....157

Getting started....157

The program....157

The header....157

Implementing Google's translation service....158

Google Translate from a linguist's perspective....159

Playing with the tool....160

Linguistic assessment of Google Translate....160

AI as a new frontier....164

Lexical field and polysemy....164

Exploring the frontier – customizing Google Translate with a Python program....166

k-nearest neighbor algorithm....167

Implementing the KNN algorithm....168

The knn_polysemy.py program....171

Implementing the KNN function in Google_Translate_Customized.py....173

Conclusions on the Google Translate customized experiment....181

The disruptive revolutionary loop....182

Summary....182

Questions....183

Further reading....183

Chapter 7: Optimizing Blockchains with Naive Bayes....186

Part I – the background to blockchain technology....187

Mining bitcoins....188

Using cryptocurrency....189

PART II – using blockchains to share information in a supply chain....190

Using blockchains in the supply chain network....193

Creating a block....194

Exploring the blocks....195

Part III – optimizing a supply chain with naive Bayes in a blockchain process....196

A naive Bayes example....196

The blockchain anticipation novelty....198

The goal – optimizing storage levels using blockchain data....199

Implementation of naive Bayes in Python....202

Gaussian naive Bayes....202

Summary....206

Questions....206

Further reading....207

Chapter 8: Solving the XOR Problem with a Feedforward Neural Network....208

The original perceptron could not solve the XOR function....209

XOR and linearly separable models....210

Linearly separable models....210

The XOR limit of a linear model, such as the original perceptron....211

Building an FNN from scratch....213

Step 1 – defining an FNN....213

Step 2 – an example of how two children can solve the XOR problem every day....214

Implementing a vintage XOR solution in Python with an FNN and backpropagation....218

A simplified version of a cost function and gradient descent....220

Linear separability was achieved....223

Applying the FNN XOR function to optimizing subsets of data....225

Summary....231

Questions....232

Further reading....232

Chapter 9: Abstract Image Classification with Convolutional Neural Networks (CNNs)....234

Introducing CNNs....235

Defining a CNN....236

Initializing the CNN....238

Adding a 2D convolution layer....239

Kernel....239

Shape....244

ReLU....244

Pooling....247

Next convolution and pooling layer....248

Flattening....249

Dense layers....249

Dense activation functions....250

Training a CNN model....250

The goal....251

Compiling the model....252

The loss function....252

The Adam optimizer....254

Metrics....255

The training dataset....255

Data augmentation....256

Loading the data....256

The testing dataset....257

Data augmentation on the testing dataset....257

Loading the data....257

Training with the classifier....258

Saving the model....259

Next steps....259

Summary....260

Questions....260

Further reading and references....260

Chapter 10: Conceptual Representation Learning....262

Generating profit with transfer learning....263

The motivation behind transfer learning....264

Inductive thinking....264

Inductive abstraction....264

The problem AI needs to solve....265

The gap concept....266

Loading the trained TensorFlow 2.x model....267

Loading and displaying the model....267

Loading the model to use it....271

Defining a strategy....274

Making the model profitable by using it for another problem....275

Domain learning....276

How to use the programs....276

The trained models used in this section....277

The trained model program....277

Gap – loaded or underloaded....278

Gap – jammed or open lanes....280

Gap datasets and subsets....282

Generalizing the (the gap conceptual dataset)....282

The motivation of conceptual representation learning metamodels applied to dimensionality....283

The curse of dimensionality....283

The blessing of dimensionality....284

Summary....285

Questions....286

Further reading....286

Chapter 11: Combining Reinforcement Learning and Deep Learning....288

Planning and scheduling today and tomorrow....289

A real-time manufacturing process....291

Amazon must expand its services to face competition....291

A real-time manufacturing revolution....292

CRLMM applied to an automated apparel manufacturing process....295

An apparel manufacturing process....296

Training the CRLMM....298

Generalizing the unit training dataset....298

Food conveyor belt processing – positive p and negative n gaps....299

Running a prediction program....303

Building the RL-DL-CRLMM....303

A circular process....304

Implementing a CNN-CRLMM to detect gaps and optimize....305

Q-learning – MDP....306

MDP inputs and outputs....307

The optimizer....310

The optimizer as a regulator....310

Finding the main target for the MDP function....313

A circular model – a stream-like system that never starts nor ends....315

Summary....320

Questions....320

Further reading....321

Chapter 12: AI and the Internet of Things (IoT)....322

The public service project....323

Setting up the RL-DL-CRLMM model....324

Applying the model of the CRLMM....326

The dataset....327

Using the trained model....329

Adding an SVM function....330

Motivation – using an SVM to increase safety levels....331

Definition of a support vector machine....332

Python function....334

Running the CRLMM....336

Finding a parking space....336

Deciding how to get to the parking lot....339

Support vector machine....340

The itinerary graph....342

The weight vector....343

Summary....344

Questions....345

Further reading....345

Chapter 13: Visualizing Networks with TensorFlow 2.x and TensorBoard....346

Exploring the output of the layers of a CNN in two steps with TensorFlow....347

Building the layers of a CNN....348

Processing the visual output of the layers of a CNN....352

Analyzing the visual output of the layers of a CNN....356

Analyzing the accuracy of a CNN using TensorBoard....363

Getting started with Google Colaboratory....363

Defining and training the model....365

Introducing some of the measurements....368

Summary....370

Questions....371

Further reading....371

Chapter 14: Preparing the Input of Chatbots with Restricted Boltzmann Machines (RBMs) and Principal Component Analysis (PCA)....372

Defining basic terms and goals....373

Introducing and building an RBM....374

The architecture of an RBM....375

An energy-based model....376

Building the RBM in Python....379

Creating a class and the structure of the RBM....379

Creating a training function in the RBM class....379

Computing the hidden units in the training function....380

Random sampling of the hidden units for the reconstruction and contractive divergence....381

Reconstruction....382

Contrastive divergence....383

Error and energy function....383

Running the epochs and analyzing the results....384

Using the weights of an RBM as feature vectors for PCA....386

Understanding PCA....391

Mathematical explanation....392

Using TensorFlow's Embedding Projector to represent PCA....396

Analyzing the PCA to obtain input entry points for a chatbot....399

Summary....401

Questions....402

Further reading....402

Chapter 15: Setting Up a Cognitive NLP UICUI Chatbot....404

Basic concepts....405

Defining NLU....405

Why do we call chatbots "agents"?....405

Creating an agent to understand Dialogflow....406

Entities....407

Intents....411

Context....416

Adding fulfillment functionality to an agent....421

Defining fulfillment....422

Enhancing the cogfilmdr agent with a fulfillment webhook....423

Getting the bot to work on your website....426

Machine learning agents....427

Using machine learning in a chatbot....427

Speech-to-text....427

Text-to-speech....428

Spelling....430

Why are these machine learning algorithms important?....432

Summary....433

Questions....434

Further reading....434

Chapter 16: Improve the Emotional Intelligence Deficiencies of Chatbots....436

From reacting to emotions, to creating emotions....437

Solving the problems of emotional polysemy....437

The greetings problem example....438

The affirmation example....439

The speech recognition fallacy....439

The facial analysis fallacy....440

Small talk....441

Courtesy....441

Emotions....444

Data logging....444

Creating emotions....447

RNN research for future automatic dialog generation....452

RNNs at work....453

RNN, LSTM, and vanishing gradients....454

Text generation with an RNN....455

Vectorizing the text....455

Building the model....456

Generating text....458

Summary....460

Questions....461

Further reading....461

Chapter 17: Genetic Algorithms in Hybrid Neural Networks....462

Understanding evolutionary algorithms....463

Heredity in humans....463

Our cells....464

How heredity works....464

Evolutionary algorithms....465

Going from a biological model to an algorithm....466

Basic concepts....466

Building a genetic algorithm in Python....469

Importing the libraries....469

Calling the algorithm....470

The main function....470

The parent generation process....471

Generating a parent....471

Fitness....472

Display parent....473

Crossover and mutation....474

Producing generations of children....476

Summary code....479

Unspecified target to optimize the architecture of a neural network with a genetic algorithm....480

A physical neural network....480

What is the nature of this mysterious S-FNN?....481

Calling the algorithm cell....482

Fitness cell....483

ga_main() cell....484

Artificial hybrid neural networks....485

Building the LSTM....486

The goal of the model....487

Summary....488

Questions....489

Further reading....489

Chapter 18: Neuromorphic Computing....490

Neuromorphic computing....491

Getting started with Nengo....492

Installing Nengo and Nengo GUI....493

Creating a Python program....495

A Nengo ensemble....495

Nengo neuron types....496

Nengo neuron dimensions....497

A Nengo node....497

Connecting Nengo objects....499

Visualizing data....499

Probes....504

Applying Nengo's unique approach to critical AI research areas....508

Summary....511

Questions....512

References....512

Further reading....512

Chapter 19: Quantum Computing....514

The rising power of quantum computers....515

Quantum computer speed....516

Defining a qubit....519

Representing a qubit....519

The position of a qubit....520

Radians, degrees, and rotations....521

The Bloch sphere....522

Composing a quantum score....523

Quantum gates with Quirk....523

A quantum computer score with Quirk....525

A quantum computer score with IBM Q....526

A thinking quantum computer....529

Representing our mind's concepts....529

Expanding MindX's conceptual representations....529

The MindX experiment....530

Preparing the data....530

Transformation functions – the situation function....530

Transformation functions – the quantum function....533

Creating and running the score....533

Using the output....535

Summary....536

Questions....536

Further reading....537

Appendix: Answers to the Questions....538

Chapter 1 – Getting Started with Next-Generation Artificial Intelligence through Reinforcement Learning....538

Chapter 2 – Building a Reward Matrix – Designing Your Datasets....540

Chapter 3 – Machine Intelligence – Evaluation Functions and Numerical Convergence....541

Chapter 4 – Optimizing Your Solutions with K-Means Clustering....542

Chapter 5 – How to Use Decision Trees to Enhance K-Means Clustering....544

Chapter 6 – Innovating AI with Google Translate....545

Chapter 7 – Optimizing Blockchains with Naive Bayes....547

Chapter 8 – Solving the XOR Problem with a Feedforward Neural Network....548

Chapter 9 – Abstract Image Classification with Convolutional Neural Networks (CNNs)....550

Chapter 10 – Conceptual Representation Learning....551

Chapter 11 – Combining Reinforcement Learning and Deep Learning....553

Chapter 12 – AI and the Internet of Things....554

Chapter 13 – Visualizing Networks with TensorFlow 2.x and TensorBoard....556

Chapter 14 – Preparing the Input of Chatbots with Restricted Boltzmann Machines (RBMs) and Principal Component Analysis (PCA)....557

Chapter 15 – Setting Up a Cognitive NLP UICUI Chatbot....558

Chapter 16 – Improve the Emotional Intelligence Deficiencies of Chatbots....559

Chapter 17 – Genetic Algorithms in Hybrid Neural Networks....560

Chapter 18 – Neuromorphic Computing....561

Chapter 19 – Quantum Computing....563

Other Books You May Enjoy....566

Index....570

Understand the fundamentals and develop your own AI solutions in this updated edition packed with many new examples

Key Features

• AI-based examples to guide you in designing and implementing machine intelligence
• Build machine intelligence from scratch using artificial intelligence examples
• Develop machine intelligence from scratch using real artificial intelligence

Book Description

AI has the potential to replicate humans in every field. Artificial Intelligence By Example, Second Edition serves as a starting point for you to understand how AI is built, with the help of intriguing and exciting examples.

This book will make you an adaptive thinker and help you apply concepts to real-world scenarios. Using some of the most interesting AI examples, right from computer programs such as a simple chess engine to cognitive chatbots, you will learn how to tackle the machine you are competing with. You will study some of the most advanced machine learning models, understand how to apply AI to blockchain and Internet of Things (IoT), and develop emotional quotient in chatbots using neural networks such as recurrent neural networks (RNNs) and convolutional neural networks (CNNs).

This edition also has new examples for hybrid neural networks, combining reinforcement learning (RL) and deep learning (DL), chained algorithms, combining unsupervised learning with decision trees, random forests, combining DL and genetic algorithms, conversational user interfaces (CUI) for chatbots, neuromorphic computing, and quantum computing.

By the end of this book, you will understand the fundamentals of AI and have worked through a number of examples that will help you develop your AI solutions.

What you will learn

• Apply k-nearest neighbors (KNN) to language translations and explore the opportunities in Google Translate
• Understand chained algorithms combining unsupervised learning with decision trees
• Solve the XOR problem with feedforward neural networks (FNN) and build its architecture to represent a data flow graph
• Learn about meta learning models with hybrid neural networks
• Create a chatbot and optimize its emotional intelligence deficiencies with tools such as Small Talk and data logging
• Building conversational user interfaces (CUI) for chatbots
• Writing genetic algorithms that optimize deep learning neural networks
• Build quantum computing circuits

Who this book is for

Developers and those interested in AI, who want to understand the fundamentals of Artificial Intelligence and implement them practically. Prior experience with Python programming and statistical knowledge is essential to make the most out of this book.