Artificial Intuition : The Improbable Deep Learning Revolution
I challenge you to find a study that is as fascinating and exciting as Deep Learning. Deep Learning is a new kind of Artificial Intelligence technology that emerged at around 2012. Daniel Kahneman, the Nobel prize-winning psychologist, and economist introduced the idea that the human brain has two kinds of cognitive or thinking machinery. He dubbed this "Thinking Fast and Slow". The fast cognitive mechanism is commonly understood as intuition. The slower kind is of the deliberative kind and is understood as rational thought. This book frames the emergent technology of Deep Learning as a kind of artificial intuition. This idea of Deep Learning as intuition is an extremely compelling narrative and can form a very solid mental model to think about its ramifications. We are all going to be profoundly impacted by this new kind of Artificial Intelligence and it is critical we all develop at least a good intuition of how and why it will transform our technology and ultimately our civilization.
The images on the cover of the book are generated by Deep Learning automation.
Preface 12
1 Introduction 14
A New Kind of Artificial Intelligence 16
What is Deep Learning? 20
Explaining to a Five Year Old 23
Tribes of Artificial Intelligence 24
The Deep Learning Industrial Complex 31
Heart Throbbing Hardware 36
The Sputnik Moment in Asia 40
2 Deep Learning is Eating the World 44
Self Driving Machines 44
Artistic Machines 47
The Uncanny Valley 49
Reading Human Minds 51
Machines that Teach 54
Summary 55
3 Our Cognitive Stack 59
Variety of Human Intelligences 60
Learning using Intuition 64
Natural Stupidity 71
Hacking Intuition 74
Human Legacy Bias 77
4 Artificial Intuition 82
Limits of Rationality 83
Artificial Intuition 86
Characteristics of Artificial Intuition 93
Growing Innate Intuition 97
Intuition Demolishes Logic 101
5 Uncertain Reality 105
Unknowable Knowns 106
Probability and Causality 108
Cargo Cult Science and Alchemy 116
Exploitation, Exploration and Representation 124
A Roadmap for Deep Learning 129
Meta Model 134
Stochastic Gradient Descent 138
A Reality Checklist 145
Deep Learning Limitations 147
Artificial General Intelligence 153
6 The Learning Universe 156
Alien Intelligences in Our Midst 157
Biological Brains are Digital 160
The Holographic Principle 163
Imaginary Numbers 169
Non-Equilibrium Information Dynamics 175
Chaos and Entanglement 183
Deconstructing Complexity 188
Sand Piles 194
The Origins of Life and Learning 197
7 Capitalism in the Age of Intelligence 203
Disruption with Learning Platforms 204
Architectures of Participation 209
Levels of Automation 212
The Responsive Corporation 215
8 Knowledge Creation 219
Accelerated Productivity 220
The Agile Manifesto 223
Pattern Languages 227
Alchemy and Black Magic 232
Data is the new Vineyard 234
Humans in the Loop 235
9 Contextual Adaptation 240
Biologically Inspired Architecture 241
Generalization 243
Deep Teaching 247
Three Dimensions of Cognition 250
The Spectrum of Embodied Learning 256
10 Conversational Cognition 262
Coordinating Rationality and Intuition 265
Learning to Communicate 270
The Many Body Problem 274
Explainability 277
AlphaGo Zero’s Self-Learning 279
Architectures of Collaboration 287
Machine Self-Awareness 293
The Strange Loop 298
11 Human Compatible AI 304
Jobs that are Safe 305
Temporal Impedance Mismatch 311
Decisions and Responsibility 312
AI Regulation 314
Seven Deadly Sins 317
Provably Beneficial AI 321