Exploring the three key disciplines of intelligent systems
As artificial intelligence (AI) and machine learning technology continue to develop and find new applications, advances in this field have generally been focused on the development of isolated software data analysis systems or of control systems for robots and other devices. By applying model-based systems engineering to AI, however, engineers can design complex systems that rely on AI-based components, resulting in larger, more complex intelligent systems that successfully integrate humans and AI.
Engineering Intelligent Systems relies on Dr. Barclay R. Brown’s 25 years of experience in software and systems engineering to propose an integrated perspective to the challenges and opportunities in the use of artificial intelligence to create better technological and business systems. While most recent research on the topic has focused on adapting and improving algorithms and devices, this book puts forth the innovative idea of transforming the systems in our lives, our societies, and our businesses into intelligent systems. At its heart, this book is about how to combine systems engineering and systems thinking with the newest technologies to design increasingly intelligent systems.
Engineering Intelligent Systems readers will also find: - An introduction to the fields of artificial intelligence with machine learning, model-based systems engineering (MBSE), and systems thinking - the key disciplines for making systems smarter - An example of how to build a deep neural network in a spreadsheet, with no code or specialized mathematics required - An approach to the visual representation of systems, using techniques from moviemaking, storytelling, visual systems design, and model-based systems engineering - An analysis of the potential ability of computers to think, understand and become conscious and its implications for artificial intelligence - Tools to allow for easier collaboration and communication among developers and engineers, allowing for better understanding between stakeholders, and creating a faster development cycle - A systems thinking approach to people systems - systems that consist only of people and which form the basis for our organizations, communities and society
Engineering Intelligent Systems offers an intriguing new approach to making systems more intelligent using artificial intelligence, machine learning, systems thinking, and system modeling and therefore will be of interest to all engineers and business professionals, particularly systems engineers.
Table of Contents
Acknowledgments xi
Introduction xiii
Part I Systems and Artificial Intelligence 1
1 Artificial Intelligence, Science Fiction, and Fear 3
1.1 The Danger of AI 3
1.2 The Human Analogy 5
1.3 The Systems Analogy 6
1.4 Killer Robots 7
1.5 Watching the Watchers 9
1.6 Cybersecurity in a World of Fallible Humans 12
1.7 Imagining Failure 17
1.8 The New Role of Data: The Green School Bus Problem 23
1.9 Data Requirements 25
1.9.1 Diversity 26
1.9.2 Augmentation 28
1.9.3 Distribution 29
1.9.4 Synthesis 30
1.10 The Data Lifecycle 31
1.11 AI Systems and People Systems 41
1.12 Making an AI as Safe as a Human 45
References 48
2 We Live in a World of Systems 49
2.1 What Is a System? 49
2.2 Natural Systems 51
2.3 Engineered Systems 53
2.4 Human Activity Systems 54
2.5 Systems as a Profession 54
2.5.1 Systems Engineering 54
2.5.2 Systems Science 55
2.5.3 Systems Thinking 55
2.6 A Biological Analogy 56
2.7 Emergent Behavior: What Makes a System, a System 56
2.8 Hierarchy in Systems 60
2.9 Systems Engineering 64
3 The Intelligence in the System: How Artificial Intelligence
Really Works 71
3.1 What Is Artificial Intelligence? 71
3.1.1 Myth 1: AI SystemsWork Just Like the Brain Does 72
3.1.2 Myth 2: As Neural Networks Grow in Size and Speed, They Get Smarter 72
3.1.3 Myth 3: Solving a Hard or Complex Problem Shows That an AI Is Nearing Human Intelligence 73
3.2 Training the Deep Neural Network 75
3.3 Testing the Neural Network 76
3.4 Annie Learns to Identify Dogs 76
3.5 How Does a Neural NetworkWork? 80
3.6 Features: Latent and Otherwise 81
3.7 Recommending Movies 82
3.8 The One-Page Deep Neural Network 84
4 Intelligent Systems and the People they Love 97
4.1 Can Machines Think? 97
4.2 Human Intelligence vs. Computer Intelligence 98
4.3 The Chinese Room: Understanding, Intentionality, and Consciousness 99
4.4 Objections to the Chinese Room Argument 104
4.4.1 The Systems Reply to the CRA 104
4.4.2 The Robot Reply 104
4.4.3 The Brain Simulator Reply 105
4.5 Agreement on the CRA 107
4.5.1 Analyzing the Systems Reply: Can the Room Understand when Searle Does Not? 109
4.6 Implementation of the Chinese Room System 114
4.7 Is There a Chinese-Understanding Mind in the Room? 115
4.7.1 Searle and Block on Whether the Chinese Room Can Understand 116
4.8 Chinese Room: Simulator or an Artificial Mind? 118
4.8.1 Searle on Strong AI Motivations 120
4.8.2 Understanding and Simulation 121
4.9 The Mind of the Programmer 127
4.10 Conclusion 133
References 135
Part II Systems Engineering for Intelligent Systems 137
5 Designing Systems by Drawing Pictures and Telling Stories 139
5.1 Requirements and Stories 139
5.2 Stories and Pictures: A Better Way 141
5.3 How Systems Come to Be 141
5.4 The Paradox of Cost Avoidance 145
5.5 Communication and Creativity in Engineering 147
5.6 Seeing the Real Needs 148
5.7 Telling Stories 150
5.8 Bringing a Movie to Life 153
5.9 Telling System Stories and the Combination Pitch 157
5.10 The Combination Pitch 159
5.11 Stories in Time 160
5.12 Roles and Personas 161
6 Use Cases: The Superpower of Systems Engineering 165
6.1 The Main Purpose of Systems Engineering 165
6.2 Getting the Requirements Right: A Parable 166
6.2.1 A Parable of Systems Engineering 168
6.3 Building a Home: A Journey of Requirements and Design 170
6.4 Where Requirements Come From and a Koan 173
6.4.1 A Requirements Koan 177
6.5 The Magic of Use Cases 177
6.6 The Essence of a Use Case 181
6.7 Use Case vs. Functions: A Parable 184
6.8 Identifying Actors 186
6.8.1 Actors Are Outside the System 187
6.8.2 Actors Interact with the System 187
6.8.3 Actors Represent Roles 188
6.8.4 Finding the Real Actors 188
6.8.5 Identifying Nonhuman Actors 191
6.8.6 DoWe Have ALL the Actors? 193
6.9 Identifying Use Cases 193
6.10 Use Case Flows of Events 196
6.10.1 BalancingWork Up-Front with Speed 199
6.10.2 Use Case Flows and Scenarios 201
6.10.3 Writing Alternate Flows 202
6.10.4 Include and Extend with Use Cases 203
6.11 Examples of Use Cases 205
6.11.1 Example Use Case 1: Request Customer Service from Acme Library Support 205
6.11.2 Example Use Case 2: Ensure Network Stability 206
6.11.3 Example Use Case 3: Search for Boat in Inventory 206
6.12 Use Cases with Human Activity Systems 207
6.13 Use Cases as a Superpower 208
References 208
7 Picturing Systems with Model Based Systems Engineering 209
7.1 How Humans Build Things 209
7.2 C: Context 212
7.2.1 Actors for the VX 213
7.2.2 Actors for the Home System 216
7.3 U: Usage 217
7.4 S: States and Modes 221
7.5 T: Timing 224
7.6 A: Architecture 225
7.7 R: Realization 230
7.8 D: Decomposition 234
7.9 Conclusion 238
8 A Time for Timeboxes and the Use of Usage Processes 239
8.1 Problems in Time Modeling: Concurrency, False Precision, and Uncertainty 240
8.1.1 Concurrency 240
8.1.2 False Precision 240
8.1.3 Uncertainty 241
8.2 Processes and Use Cases 242
8.3 Modeling: Two Paradigms 243
8.3.1 The Key Observation 244
8.3.2 Source of the Problem 246
8.4 Process and System Paradigms 247
8.5 A Closer Examination of Time 248
8.6 The Need for a New Approach 251
8.7 The Timebox 252
8.8 Timeboxes with Timelines 257
8.8.1 Thinking in Timeboxes 257
8.9 The Usage Process 258
8.10 Pilot Project Examples 262
8.10.1 Pilot Project: The Hunt for Red October 262
8.10.2 Pilot Project: FAA 265
8.10.3 Pilot Project: IBM Agile Process 267
8.11 Summary: A New Paradigm Modeling Approach 269
8.11.1 The Impact of New Paradigm Models 270
8.11.2 The Future of New Paradigm Models 271
References 272
Part III Systems Thinking for Intelligent Systems 275
9 Solving Hard Problems with Systems Thinking 277
9.1 Human Activity Systems and Systems Thinking 277
9.2 The Central Insight of Systems Thinking 279
9.3 Solving Problems with Systems Thinking 281
9.3.1 Identify a Problem 281
9.3.2 Find the Real Problem 282
9.3.3 Identify the System 284
9.4 Understanding the System 285
9.4.1 Rocks Are Hard 288
9.4.2 Heart and Soul 290
9.4.3 Confusing Cause and Effect 292
9.4.4 Logical Fallacies 296
9.5 System Archetypes 298
9.5.1 Tragedy of the Commons 299
9.5.2 The Rich Get Richer 300
9.6 Intervening in a System 302
9.7 Testing Implementing Intervention Incrementally 315
9.8 Systems Thinking and theWorld 316
10 People Systems: A New Way to Understand the World 317
10.1 Reviewing Types of Systems 317
10.2 People Systems 318
10.3 People Systems and Psychology 320
10.4 Endowment Effect 323
10.5 Anchoring 324
10.6 Functional Architecture of a Person 325
10.7 Example: The Problem of Pollution 327
10.8 Speech Acts 332
10.8.1 People System Archetypes 337
10.8.1.1 Demand Slowing 339
10.8.1.2 Customer Service 340
10.9 Seeking Quality 341
10.10 Job Hunting as a People System 344
10.10.1 Who Are You? 345
10.10.2 What Do You Want to Do? 345
10.10.3 For Whom? 347
10.10.4 Pick a Few 348
10.10.5 Go Straight to the Hiring Manager 349
10.10.6 Follow Through 351
10.10.7 Broaden Your View 352
10.10.8 Step Two 352
10.11 Shared Service Monopolies 354
References 356
Index 357
